dev_rc_branch

[reef bobcat]

Has the official side provided any method to pass data from JS to SQF, or is it still under development?

This should work.

[slate tide]

There is no need to use DLLs. Data can be sent from JS to SQF and from SQF to JS. It seems that a sqf file can be called using the Arma 3 JS API if necessary. Furthermore, the use of DLLs is not necessary.

Read the documentation; it explains the possible relationships.

By the way, where did you get the information about the DLLs?

[reef bobcat]

Yes, I saw it, which indicates that the related functionality is currently under development.

[reef bobcat]

https://community.bistudio.com/wiki/callExtension

Look at this document.

[full sonnet]

True. Might be cooler if it could be embedded into missions, without mod.
But thats so annoying to do as a developer

You can display basically one HTML page.
You have no other internet access, its running in a isolated very restricted sandbox.
You can design the page however you want, and can run (basically) any javascript to make it interactive. And communicate with SQF which lets you load files from a PBO.

So everything you can put into one html page, or put into a PBO. You can do.

Just no external internet access. No microphone/camera stuff. No browser local storage. And some other restrictions to make it "safe"

[full sonnet]

Its not, its all "done" and ready to use

[reef bobcat]

The JavaScript API shows how to load game images in JavaScript, but how can data be returned from SQF? For example, numbers or boolean values?

[slate tide]

How do you create a .DLL extension? Is there any documentation available? Because I have quite a few ideas.

[full sonnet]

External websites/internet access are disabled.
The feature exists and works, but we disabled it for security.

[slate tide]

logic I understand

[full sonnet]

SQF can ExecJS to execute javascript in the browser. You can pass any data along you want.
Javascript can run an "alert", which lets it send any string to SQF. You can pass any data you want (There is some size restriction)

Both of those methods are also how you can load game textures.
An alert sends a request to the game. The game sends base64 encoded string version of the image back by executing javascript.

[full sonnet]

Just read the Extensions wiki page, should have everything

[slate tide]

I never found it, I think I'm a bit stupid x')

x')

i found it

i'm stupid x')

[slate tide]

Are there any restrictions in the DLLs?

[reef bobcat]

alert is just a prompt, right? I mean passing data to an SQF variable.

[full sonnet]

To use it with BattlEye, BE needs to approve it.
If you don't use BattlEye, no.

[full sonnet]

it just passes the string to a SQF eventhandler

[reef bobcat]

Okay, I understand.

[solid crown]

Just curious in general: What's the reason behind the JS alert being used as the "channel" to return data to the game/SQF?

[tepid dock]

If after reading the wiki you still have some dll-related questions, then feel free to ask them in #arma3_tools 🙂

[feral tangle]

Did I hallucinate that there is a map = helipad named property? 🤔 😵‍💫

[grim lion]

@tepid dock Here's the OS that I created that you mentioned above: https://github.com/InnovativeDevSolutions/os

[near carbon]

My upcoming game mode uses @grim lion JS Stuff + some my own

Our Capture UI for OBJs uses it, so much better

Love the js stuff

[grim lion]

Just released the os on the workshop

[full sonnet]

It exists.

[full sonnet]

Dear dev diary, today I wrote bad code. But its fine, I left a note
Just don't have hundreds of aircraft in MP I guess

[verbal venture]

You should make a video, like people did with the cs and tf2 source code when they leaked.
"The sanity of BI developers"

[unreal lava]

Only makes me wanna try having hundreds of aircraft in MP now

[gray quarry]

that'll be fun, maybe one day the arma channel could do that!

arma code comments without context

few rounds, preferably long form videos, 10 min for the milking xd

might be too embarrassing

[teal elk]

Hi guys, anyone have by chance the proyect that some did in the past of an local, offline wiki ?

i have a discord bot that access to the biki, but with the cloudflare thingy...

[viscid onyx]

#community_wiki message

[teal elk]

tanku ❤️

[teal elk]

someone tried to do a MCP with the wiki?

[late harbor]

Maybe this is useful?
https://github.com/acemod/arma3-wiki

[teal elk]

Well ive used the biki files to create a DB, so i tryed to made a mcp lol
https://github.com/Flyingtarta/MCP_ARMA3_BIKI/tree/main/arma3biki_mcp

and when i say "i tryed" was actually claude-code

[full sonnet]

Note while Windows 7 support was already canned. Windows 8 and 8.1 support is also coming up.
We already bumped the min requirement to windows 10, but so far only definitively dropped windows 7.

Not in 2.22 yet, but Microsoft removed windows 8.1 support from their latest compiler, which we will likely upgrade towards later this year.
So if you're running Windows 8.1 still (We all know no-one would be running windows 7 anymore 🤣 ), consider preparing for an upgrade.

This ALSO includes dropping Windows Server 2008 and 2012.

Someone could write a wrapper that backports new API to older OS, and if that is present the game would still run.
Such a project probably already exists somewhere on GitHub. But it'll be too much effort 😄

[unreal arrow]

Considering RAM prices are about to skyrocket, old PCs are there to stay

[unreal arrow]

Not fluent in CSAT but I think he says &@&#%!!&£&& GRENADE!

[gray quarry]

hm, should vxkex work fine then?

with or without battleye tho?

[full sonnet]

I don't know what that is

[tall pumice]

i just remembered.. does anyone else have broken mod logos on dev-branch or is it just me?

my ws and rf files both dont show the icons in virtual arsenal on some items, also the icon change when hovering over it (to the colored variant) doesn't seem to be working

example:

the icon works on the bandana since it's using a vanilla model file, i guess

the ws mask uses our own model file, which isn't detected as part of the addon

and if i move the mouse onto the white box in the bottom right, a popup with "any" something something image can't be found shows up

[gray quarry]

someone's project using image file execution options feature of windows 7 and later to extend kernel via dlls

[tall pumice]

ok, for the fun of it i just reverted to profiling branch and there the mod logos are all showing up correctly.. so whatever is going on, it's a dev-branch only issue i guess

[waxen knot]

Seems OK? Icons show up fine in Arsenal on devbranch even with diagnostics exe.

[tall pumice]

however, the icon change on mouse-over doesnt work on profiling either

hm ok, i see. the mouse-over icon change works if i actually load the official dlc files via the launcher (dlc menu). if i load it as a regular mod from local files, the mouse-over doesn't work

[full sonnet]

Well if it says "any" then probably script error?
Nil to string?

[full sonnet]

But why would it work here

[tall pumice]

but it works on the other item

and it works in the main menu, that small mod list

[full sonnet]

Well its script that pulls the info of the item. Its just failing on it but what exactly is afiling

[tall pumice]

ok, i reverted to dev-branch again and used the dlc menu to enable ws and rf. there it works fine as well.

however -- if i use the workshop mod (compatibility data), the icons are also broken

[ebon quail]

[tall pumice]

tl;dr - the issue seems to be with mod handling, not the official launcher cdlc handling

the mod.cpp files are identical. the mod config entry is the same as well of course

mod file has a meta.cpp that is auto-generated by steam i believe. other than that, the only difference is the folder name

i'm 100% sure this is a new issue, though. i never noticed it in the past

actually, yeah. if profiling has no data changes, but it works there (mostly, except the mouse-over), but on dev-branch it is broken, then there must have been some script changes that are causing it now.

[ebon quail]

cdlc depots have the same manifests atleast on dev/prof so

[full sonnet]

The DLC menu in launcher also only sets -mod parameter.
meta.cpp is only used in 4 places.
Crash reporting, server modlist, getLoadedModsInfo script command the workshop itemID.

So I don't think that should matter.
And the rest of the data is the same.
Only leaves the folder name?

You could try to check the classnames with this
https://github.com/acemod/ACE3/blob/master/addons/common/functions/fnc_getDLC.sqf
getAssetDLCInfo on the model.

Ah found it

There was a arsenal change in november.

[internal] Changed logic of detecting mod, which it belongs - for DLC assets there is now another, special check which utilizes getAssetDLCInfo. If model belongs to DLC, then it will be marked as so - even if i.e. class was defined in another mod.

BIS_fnc_getDLC was added. Parameter is the items config class.
That is probably failing now and returning empty string

I don't have time to look deeper. But testing through what that function does should find the problem

[tall pumice]

at least that seems fine

ok, without looking at the arsenal code, i'm pretty sure it's using "ws" as path, because that's also the addon folder name

and since compat data, etc. doesn't have that folder name, it's failing in arsenal to fetch the correct data

or maybe not. looking at the script right now it doesn't seem like it. _dlc = _itemCfg call bis_fnc_getDLC; and then a bit further down _ctrlDLC ctrlseteventhandler [ "buttonclick", format ["uiNamespace setvariable ['RscDisplayDLCPreview_dlc','%1']; ctrlparent (_this select 0) createDisplay 'RscDisplayDLCPreview';",_dlc] ];

ok, there's more code. also a macro. but that's where it stops for me. i'm not a good enough programmer to investigate this 😄

[full sonnet]

The old getDLC code was this

    {
        private _dlc = "";
        private _addons = configsourceaddonlist _this;
        if (count _addons > 0) then {
            private _mods = configsourcemodlist (configfile >> "CfgPatches" >> _addons select 0);
            if (count _mods > 0) then {
                _dlc = _mods select 0;
            };
        };
        _dlc
    }

Parameter is also config, I'd assume that also returned "ws"

The icon then comes from
modParams [_dlcName,["logo"]]

so modParams ["ws",["logo"]] ? That work for both variants?
But that part of the code hasn't changed in quite a while

[full sonnet]

You have a 32-bit operating system and you can't run the 64-bit version of the game.
just found this, lol

[full sonnet]

Finally we have space

[verbal venture]

diag option 👀

[full sonnet]

oh right, I forgot that

[verbal venture]

no more renaming files when I want to render an ocap map?

[gray quarry]

finally about time

why is server listed if it's not necessary since -server exists?

[full sonnet]

Because headless client

or does -client work on client binary?
But you also want the no steam, and no battleye, and stuff

[gray quarry]

headless client also can be launched with game exe passing -client to it

[full sonnet]

The answer is, because

[gray quarry]

pointless since the launcher doesn't work properly without steam

[full sonnet]

launching multiple instances bla bla

[gray quarry]

game exe can be used for that no problem

launcher properly handles -server by not launching the battleye launcher when be is enabled

also on the topic of launcher changes, can we remove the screen that blocks the user from accessing the mod list and do operations within it?

and how about a way to copy the full launch arguments? so one doesn't need to launch at least once to grab it from the rpt

[gray quarry]

this makes sense to do with multiple instances in mind, eg different modlist or removing client only mods
and also why not add a way to make servermod sets right from the launcher while we're at it?

[verbal venture]

because not every server has the same setup?
for me, I would need to prepend "mods/" before every entry
and someone uses full names as seen on workshop, someone uses IDs
There's just too much variety to get something useful

[gray quarry]

that would be the point of using the launcher for modlists

[tall pumice]

funny bug with glasses transparency. left side is correct, right side is bug. it works on most vanilla uniforms, but not all of them. it's mostly broken on contact and laws of war clothes. also noticed this in pretty much all of our new ws and rf uniforms, so i started to do some digging. for some reason, i can fix the issue on one of our new uniforms if i am adding a "nvg" proxy to lod0. however, if i am trying to fix our djellaba clothes with the same thing, it does not work. even if i straight up copy&paste all proxies over, it does not work. also curious.. why nvg proxy? is this thing even used? half the uniforms don't have that, they use the hmd proxy.

basically all facewear with transparency is affected

truth told, it might as well the right side that is correct, and the left is the bugged one. no clue what the wanted effect was. i'm just guessing left side is the correct one because that's how the combat goggles etc. are looking on generic nato uniforms... the default setup, kinda

[feral tangle]

I swear everytime I look in here, Lex is breaking something. 🙄

[tall pumice]

i'm mostly just posting here when finding bugs :p

back in the days i would post on the forum, but those days are long gone

also funny, some old uniforms have a "radio" and "backpack2" proxy. maybe radio item was supposed to be visible on the uniform at some point?

now i fixed it on another uniform with replacing all proxies.... does alpha sorting apply to them?

[full sonnet]

Won't fix

[tall pumice]

well, i will fix it in my stuff if i can, that's good enough for me. just figured i point it out

what's weird is, it doesn't appear to be a rendering issue. at least not of the kind we know from the game. it just seems the transparency is stronger in one model than in the other

[tall pumice]

this is the proxy setup, if anyone wants to test it. fixed all of my uniforms with it now. lots of seemingly useless crap in there, but that's what's in the vanilla uniforms where it works

i just had a funny thought. what if transparency itself is not "broken" but because of duplicated(?) proxies, the model gets rendered multiple times, and thus the transparency appears darker 😄

i have no idea, just making shit up. this issue is really weird. but it's noticeable that this only seems to pop up on more recent (dlc) uniforms, where someone apparently cleaned up the proxy list.

[unreal lava]

you'd see z-fighting in that case, unless there's also even the slightest amount of offset between the two 🤷

[tall pumice]

yeah, probably. i dunno. it's just odd that it is clearly proxy related

[full sonnet]

Mh actually, yeah why do mods not show that they have extensions up there.
Well now they do 🤷

[gray quarry]

great addition about time :3

[unreal arrow]

These are the options: ```
uavControl [uav, "driver"]
uavControl [uav, "gunner"]
uavControl [uav, "crew"]
uavControl [uav, "all"]
uavControl [uav, "allbutcrew"]
uavControl [uav, "allbutdead"]

[full sonnet]

Found a probably decade old bug in publisher 😄
When you have extensions. But only one version 64 or 32, and not the other.
Its supposed to tell you that you're missing the other variant.

But there is a typo in the code.
The 32bit warning only shows, if you have 32bit extensions in the mod.
But the warning is to tell you its missing, so, you don't have. So the warning never shows 😢

Oh well, too late to fix now XD Though I guess I am fixing it, atleast the 64-bit missing warning

[pallid cliff]

Add a warning if there is only a 32 bit extension😁

[ebon quail]

error*

[full sonnet]

Yeah exactly. That was already there, it just never worked. So I'll just fix that one. And remove the "you're missing 32-bit" one

Probably no-one has ever seen these error's here.
The only others were if the DLL is corrupted

Probably also that

Also the tooltips take like half a second of hovering before they show up. By that time I've already given up and accepted that there isn't any tooltip
Launcher was doing the same, so annoying

[pallid cliff]

I hate it when tooltips have a delay

[full sonnet]

I found unfinished content 😄
There is a size display, telling you how large your "new" version is, compared to the version that is currently on the workshop.

[verbal venture]

(also fix the freeze when uploading large mod, thx)

[full sonnet]

Also these are a thing 😄

[full sonnet]

No repro, can't fix

[verbal venture]

it froze for me almost every time I uploaded something

[full sonnet]

This looks better to me.
This all used to be there once, but was removed when 64-bit support was added and I cannot find a reason why

You now know how old your mod is and when the last update was. And a nice plausibility check on the size difference.
It also used to list the name of the Author of the mod, but I'd think you'd know your own....
Oh.. Unless you are only a contributor and the author is someone else 🤔 But we already have the Contributor banner at the top for that

[digital vessel]

if you are looking at it, I've had problems with publisher and what I assume is a too long file path
(I have long file path registry-enabled but it had no effect, it may be something you have to also enable inside the app)

[grave python]

It freezes at the end of the first upload of this mod, But it won't when updating

[full sonnet]

what us "this"?

[pallid cliff]

Iirc it happens with every mod that gets uploaded for the first time. I can test that later when I am at home.

[full sonnet]

There you go

[pallid cliff]

What am I looking at?

[full sonnet]

I added an X button

[pallid cliff]

So you can manage mods with game running and then just start it again?

[full sonnet]

Yes. I tried to find the person who requested that above but I can't find it

[lucid totem]

maybe nobody did and it was all in your head

[full sonnet]

Can run multiple instances now.
Can launch server, then switch to client binary and start matching client with same parameters

[dark hemlock]

yaaaaay

ty dedmen, launching test servers was a pain in the ass before

[pallid cliff]

I ditched the launcher and start my sessions with hemtt instead, for exactly that limitation. Great is gone now.

[gray quarry]

that was already possible but without mod management

[gray quarry]

me?

you always could launch different mod presets from the arrow in play button

but you cant adjust mod lists themselves when the client is running

[weary niche]

Was doing some testing today and found that the newer "WeaponChanged" eventHandler doesn't fire for the Throw weapon or on any of the Throw muzzles. I'm trying to get an event for when the grenade is swapped and am not sure if this is a bug. Reloaded and Fired muzzle events do work on at least the Throw muzzles.

[gray quarry]

^

[unreal arrow]

Could you make a ticket,I thought we had it

[weary niche]

Here you are:
https://feedback.bistudio.com/T197126

[subtle pollen]

can someone please help me with my server

[halcyon moth]

This is almost certainly the wrong channel.

[grave python]

Sorry for my incorrect translation.It is when each mod is uploaded for the first time

[compact salmon]

Has anybody actually managed to make the server browser succesfully recognize and match up server mods to subscribed steam workshop mods?

[harsh orbit]

Here's the lips issue. I'm sure they weren't this bad before. (in the video it doesn't look as bad as the game at 120+ FPS)

animations issue

[full sonnet]

Yeah the low fps animations issue, isn't new

[harsh orbit]

I had seen it before too but this is the first time I saw it in this mission with 100% repro rate 😅

[viscid onyx]

I wonder if maybe the mission attaches them to something for some purpose and the sim rate isn't being updated correctly

[harsh orbit]

actually that does make sense

[harsh orbit]

yeah I'm not crazy. it's definitely much worse 😅
https://youtu.be/QRcDf3-SdHA?si=lt6-JERRC9Z0PnT_&t=217

also that AAF soldier's animation used to be fine

I wonder if it's related to the new multithreading simulation?

[unreal lava]

can confirm at least for the attachment theory - when i was making wrappers for animation sequencing i was attaching units to static invisible objects and that was happening - simulation interval does seem to matter (ergo just ended up attaching them to Game_Logic instead)

[viscid onyx]

Yes, it's normal that simulation rate affects animations - attached objects inherit the sim rate of the parent object.
What's unusual in this case is that the animations seem to be partly or inconsistently affected, and also that it's appeared in a situation where it previously wasn't happening, without (presumably) any changes to the mission.

[unreal lava]

ah

[viscid onyx]

my theory (and it's a loose theory) is that maybe the units are being attached to a low sim object, then detached, and their sim rate isn't being fully returned to normal. Or, they're being attached to a high sim object like a logic, but parts of their simulation are getting the wrong sim rate for ~reasons~.

[carmine juniper]

Actually, an animation and a tree can cause it from what I've experienced. Not sure I've made a ticket

[trim carbon]

I remember long ago I found that certain animations would end up like that when I used BIS_fnc_ambientAnim. I think that might be one of them since my notes specifically mention "standing armed" animations having the problem.

Even after stripping the code down to just starting the animation and nothing else, I would eventually end up with units doing that. Though I could never get a 100% repro and it only seemed to happen when there were a large number of units being animated.

Attaching units to a game logic did stop it, however.

[verbal venture]

diag_exportTerrainSVG doesn't escape & (and possibly others) correctly, is that something that will get no attention, if I make a FT ticket, or something that would be worth fixing?

[full sonnet]

You mean in like town names?
Easy to fix if you can tell me where they would appear

[verbal venture]

yeah, places names
for example https://steamcommunity.com/workshop/filedetails/?id=3317364155
has places R&C Tog Deponi and Falk Stop & Shop
svg lines 834114 and 834116

And they should be R&C Tog Deponi and Falk Stop & Shop

[full sonnet]

I mean the path to it in the svg.
I don't want to download a terrain and export it

[verbal venture]

I'll send it all to you in DM with files

[peak musk]

If you poke around in this part of the code base could you also remove the semicolon in the style-tag between polyline and ellipse:
Current:

<style type="text/css"><![CDATA[ polyline{ fill: none; }; ellipse{ stroke:url(#colorForest); fill: none; stroke-width: 4; } ]]></style>

Correct:

<style type="text/css"><![CDATA[ polyline{ fill: none; } ellipse{ stroke:url(#colorForest); fill: none; stroke-width: 4; } ]]></style>

It seems like a lot of parsers don't like the additional semicolon and render the trees as black circles (tested with Firefox and GIMP)

[feral tangle]

Huh, did I miss a memo about dev branch changelog going onto the wiki?

Maybe I am just blind.

[arctic fern]

From today's changelog: Tweaked: Dev / Prof / RC change logs were re-directed to Community Wiki 😉

[feral tangle]

So yes. I am blind. 😄

I tend to automatically be drawn the the "Engine" header.

[arctic fern]

Well, it was only posted to the wiki (the full content of the change log)

[waxen knot]

Fixed: Units in combat were reporting smoke throws when chemlights were thrown
Too bad there's no SentThrowingChemlight command in the RadioProtocolX classes, and no ThrowingChemlightE_1[] words so they're completely silent even in combat.
But it's still nice to see this fixed after 13 years.

[unreal arrow]

stumbled on it by accident while doing something else, I wish I could have seen that ticket 13 y ago

[ebon quail]

The "frag out" variant? 😂

[teal snow]

Damn, can the ppeffect blur in flir not be fixed?

[teal snow]

forgive me if this is written somewhere, but what actions does this command accept

[full sonnet]

not yet public

[full sonnet]

I forgot about it, it can

Don't remember if I had a repro?

[full sonnet]

#perf_prof_branch message I do have a repro and it works

[full sonnet]

Blur is fix

[lucid totem]

fixed*, I hope 😄

"bad quality image"
"use blur"
:U

[signal edge]

@full sonnet Could you fix dev version so that white retangle car-sized boxes aren't used for unknown objects? it is especially prevalent on u minh forest from rtbf (https://steamcommunity.com/sharedfiles/filedetails/?id=3454473123)

[full sonnet]

Could they fix the missing objeccts in the terrain?

[signal edge]

it isn't just that terrain, i see it in arsenal when the object doesn't exist (.p3d) or when placed by other mods that have no p3d for whatever reason

i can ask them too but it is rampant in dev version right now

[full sonnet]

It is intentional because broken/missing objects cause issues

[signal edge]

gotcha...maybe something more placeholder-y?

[feral tangle]

A big man sized question mark.

[tall pumice]

miller head

[lucid totem]

Miller T-posing, always turned towards the player – whatever the locality

[tall verge]

The gold standard

[digital vessel]

With CBA+ACE the nlaw disposable is broken on 2.21.153537
compatibleMagazines "ACE_launch_NLAW_ready_F" returns []
has magazines[] = {"NLAW_F"}; https://github.com/acemod/ACE3/blob/master/addons/disposable/CfgWeapons.hpp#L26-L27

[carmine juniper]

This is already on live, but can't see it on any changelog?

[full sonnet]

@unreal arrow

[full sonnet]

Yeah missed that. Launcher changelog is actually under the tools area ish.
Not sure how we include that in the game changelogs actually 🤔

[gray quarry]

outside of dev into stable/perf-prof?

[carmine juniper]

No, live on Dev (at least I've tested there)

[gray quarry]

ah good

when itll come to live i wonder

[unreal arrow]

and compatibleMagazines [weapon, muzzle]

[digital vessel]

compatibleMagazines ["ACE_launch_NLAW_ready_F", "ACE_launch_NLAW_ready_F"] // also []
getArray (configfile >> "CfgWeapons" >> "ACE_launch_NLAW_ready_F" >> "magazines") // ["NLAW_F"]

[dusk hinge]

Looked back in this channel after forever and saw this 👀 cant find and further info but this message and the change log of it being added in dev branch 2.21 any the wiki link from the change log goes to no where currently haha.

What actions and commands will this give us access to for pylons?

(https://community.bistudio.com/wiki/pylonAction)

Also the reason I came back to check this channel was I wanted to say you and any other devs still working to improve A3 and give us modders more commands or syntax's to work with are awesome!

I keep encountering so many new things that make life so much easier, going great ill use that only to see "since 2.22" hahaha Any idea how close it is to being ready?

[full sonnet]

What actions and commands will this give us access to for pylons?
The question is rather, what do you want it to do

currently its not much, and especially nothing at all for weapons pylons.
Open to ideas

[carmine juniper]

Targeting pod with laser designator (AFAIK Ded ever teased one)
Working fuel pod
External missile bay (Stealthy one)
What else? A soldier transport pod for Harrier? 🤣

[tall pumice]

i think you can already script working fuel pod. you need to check if the pylon is there and then you can change the fuel consumption rate of the vehicle

ok, question is where to call that script. you'd need PylonChanged EH and probably something in init

[carmine juniper]

Was about to say. We can't know it without EachFrame it

[tall pumice]

EH on object creation, check if it's vehicle / helicopter / plane, if yes, check if it has pylons, if yes, check if it has fuel pod :p

dont think you have to eachframe it

[carmine juniper]

Well, you can change pylon on the fly (no pun intended)

[tall pumice]

that's where the PylonChanged EH comes in

[viscid onyx]

Being able to force lock/unlock with that weapon, and force fire that weapon, would be handy. Not sure whether they're within the command's scope though.

[carmine juniper]

Oh that's what you mean. I meant we have to EachFrame it for now to achieve it

[viscid onyx]

Jettison would be neat too. It can be done with scripting but a cleaner option wouldn't hurt.

[full sonnet]

I was talking about the script command

[full sonnet]

Weapon pylons have no weapons, they are magazines only.
You'd need to tell the vehicle to fire its weapon

[teal snow]

I would like to be able to command fire pylons, or weapon group of pylons, regardless of if the weapons are associated to the locality of the weapon ownership and if that seat ownership exists

I’m trying to build a scripted weapon control system that allows either seat to take ownership of a group of weapons and fire it, even if not local to them, and it needs to work in mp

There are a lot of jank ways I’m trying to achieve this, there are some script firing, but it’s weird in mp, also tried swapping pylon ownership

[full sonnet]

First documentation for "Generic" Pylons is up
https://community.bistudio.com/wiki/Arma_3:_Vehicle_Loadouts#"Generic"_Pylons
https://community.bistudio.com/wiki/pylonAction

Now is the time to have ideas of what could be added.
But many things you will be thinking of, will probably be too complicated to add, so we won't.
Or simple to do via script (like the smoke pod), so we won't do that either.

[hexed canyon]

Not sure if it’s a mobile issue but the link to https://community.bistudio.com/wiki/pylonAction#Example_1 embedded in the page doesn’t seem to function correctly yet.

[full sonnet]

Oh that was a bad copy paste 😄

[unreal arrow]

Revision: 153587

[teal snow]

What are the odds of the Cfg "sensorPosition" being fixed
it was originaly introduced to allow a sensor origin that was not the eye pos however seems to not function

[full sonnet]

need more detail of what you mean

on aircraft/vehicle sensors?
When does the position ever matter?

[teal snow]

So as part of the apache mod, the radar sits above the rotor disk, allowing you to peak the radar over the horizon>
battle field radars that extend up verticly to clear obstacles, Sam sites that do the same.
vehicles with telescopic cameras

however because its on eye pos, you need to peak the whole airframe

[full sonnet]

I'll take a look tomorrow

[mossy cloak]

Just testing the dev branch on our life server to make sure nothing breaks, and I've noticed that one of the initial remoteExec's isn't actually running the function it should.

As client on join from init.sqf, it runs a remoteExec to server_fnc_queryRequest. This file grabs server/database data and remoteExec's to client_fnc_requestReceived with remoteExecutedOwner.
This works fine on profiling and stable branches, and returns to client properly.

On the dev branch the client_fnc_requestReceived never runs.
I verified the server properly sees the remoteExecutedOwner, and have verified the remoteExec to client_fnc_requestReceived returns an empty string, implying that it successfully sent it.
I have attempted to replicate the issue, however even with the exact same params, with multiple remoteExec's back and forth, I cannot replicate the issue.

I'm at a bit of a loss as to how to figure out whats going on, hopefully someone may have some insight.
Code in server_fnc_queryRequest that remoteExecs to client.

//Works, runs the function properly
_check1 = [
    _hexIcon // "autohover"
];
_check1 remoteExec ["client_fnc_requestReceived", remoteExecutedOwner];

//Works, runs the function properly
_check1 = [
    "" // manually defining ""
];
_check1 remoteExec ["client_fnc_requestReceived", remoteExecutedOwner];

//Does not work, function never runs
_check1 = [
    _title // "" - manually defining "" works,
];
_check1 remoteExec ["client_fnc_requestReceived", remoteExecutedOwner];

[halcyon moth]

So what exactly is _title?

Are you sure it's not text rather than string?

[mossy cloak]

It is 100 a string. Params as a string with only allowed types as a string
["_title","",[""]],

[mossy cloak]

It seems directly correlated with empty strings returned from a database call

[halcyon moth]

Run toArray on it.

[full sonnet]

There was a bug with empty strings that I recently introduced.
But I found it and fixed it last week

But I don't get what the difference would be between a variagble containing an empty string, and an empty string.
the code doesn't know where the string came from, it doesn't know the difference, so why would it make a difference 🤔

[full sonnet]

ah there is only one sensor position per vehicle.
But I don't knopw whats supposed to not be working about it.
The config entry is being read, and its being used (if config isn't there, it uses eyePos as you said)

I can see it being read correctly for Transport vehicles.

But "new" sensors added via SensorManagerComponent, don't support setting a position. They just always use the center of the object.
It might be desirable to make that configurable

[mossy cloak]

toArray returns an empty array for both the variable, and the empty string.
If I change the variable ex. _title = _title + "", even though its the same value, it works. I'm really shocked as to what is going on

[lucid totem]

typeName _variable = ?

[mossy cloak]

"STRING" in both cases, when it works and when it doesnt

[lucid totem]

I would have preferred to see a "TEXT" at one point 🥲

[mossy cloak]

You and me both. The code that is being run hasn't changed for years, and seems to only break with dev branch x_x

Interestingly enough, the function WILL run if I call it instead of remoteExec'ing.

[lucid totem]

#blameDedmen

[full sonnet]

I have an idea.
There are two variants of empty string

But I don't see why that would make a difference

[mossy cloak]

Your guess is better than mine

[teal snow]

I though it might have been model centre. If we can make that configurable, that would be fantastic

[full sonnet]

How did you get _title to be empty string.
If
_title = "" does not reproduce the error? Or does it?

_title = "";
[_title] remoteExec ["client_fnc_requestReceived", remoteExecutedOwner];

vs

[""] remoteExec ["client_fnc_requestReceived", remoteExecutedOwner];

?

Cannot reproduce?

TheFunction = {diag_log [diag_frameNo, _this];};

_title = ("" select [5]);
[_title] remoteExec ["TheFunction"];

This does not reproduce either

Maybe its only broken on dev-branch but already fixed internally. But the fix was quite a while ago 🤔

I'm pretty sure the difference should be

//Works, runs the function properly
_check1 = [
    ""
];
_check1 remoteExec ["client_fnc_requestReceived", remoteExecutedOwner];

//Does not work, function never runs
_check1 = [
    ("" select [5])
];
_check1 remoteExec ["client_fnc_requestReceived", remoteExecutedOwner];

That's the only thing I can think of.

[lucid totem]

after NaN, NaS (Not a String) 😄

[full sonnet]

Checked the same again on dev branch, no difference. Its working fine for me

Maybe your server/client versions do not match?
Previous dev branch version is not compatible with current version.

This empty string thing, was broken between 10-24th february.
So last dev-branch, before wednesday, was broken. But the current one isn't anymore

[mossy cloak]

Ill take another look when I'm home, however I can confirm I was on the latest dev branch. If I attempt to remoteExec to myself (myself being server), the same issue arises. That rules out a version mismatch. Attempting to remoteExec a different client has the same result

If i can find a way to downgrade my dev build to find when the issue started, it might help. I haven't tested a dev build before this one

[raven mist]

Haven't really said anything on it as I really haven't tested it outside of "my local no worky". But anyway FWIW:

• I swapped from perf to dev the same day this #dev_rc_branch message was posted at 1:26PM EST and that's when I noticed issues on our staging.
• Keeping staging, I swapped back to perf and that resolved the issue.
• Both times I was testing where client was server, although I do believe I remember Fraali saying he had same results with a dedi and client.

The database queries being performed were done with our database addon and not ExtDB3; although we haven't touched it for at least a year; outside of updating Intercept which was done roughly 6 months ago.

[raven mist]

Uhhh okay interesting...

I swapped my branch back to dev branch and am now noticing zero issues. The only thing I changed was that I updated the title entry in our database which corresponded with my player:

UPDATE my_table SET my_title=NULL WHERE my_id='76561198071959741';

Even upon swapping it back to it's prior result:

UPDATE my_table SET my_title='Ghille Warrior' WHERE my_id='76561198071959741';

It appears I am no longer experiencing any issues...

Which is really bizarre, as I believe my branch should've been the latest per the changelog message linked in my previous message and I switched my branch via Steam.

@mossy cloak you were doing a lot more testing last night, if you're able to just vibe check I am not trolling with my assessment here; would be appreciated.

Gonna feel really, really bad if we are trolling you guys and both our branches were out of date somehow. 🥹

-# Also just to toss out for you Fraali, I am still on a62222c

[full sonnet]

Could maybe look at old RPT's

and check the build number there

[raven mist]

Good thinking- here appears to be the build I was running then

Exe timestamp: 2026/03/04 01:01:27
Current time:  2026/03/04 13:26:35

Type: Public
Build: Development
Version: 2.21.153584

[mossy cloak]

Doing a bit more testing.
I would assume that that the issue stems from how empty strings are handled within the dev branch has changed, and that the addon we have based on intercept is having some type of issue with how the data is accessed natively?
Database query SELECT '' is being run, and returns properly. Any remoteExec(/Call) that attempts to pass this value provided silently fails, although every other command seems to work just fine like call, spawn etc. the value seems to behave identical to normal strings, so I wonder if maybe due to a recent optimization of how empty strings are stored not matching up properly? Just theories, I don't deal with C/C++ very often.
If there is any way I can potentially scrape memory values/create dumps for this, I (probably) have the ability to do so.
ProductVersion - ["Arma 3","Arma3",221,153584,"Development",true,"Windows","x64","development"]
Image of issue/results.
https://i.imgur.com/uYbBCSO.png
Code Run for testing

o_localTest = nil;
o_localChecking = nil;
o_remoteExecTest1 = nil;
o_remoteExecTest2 = nil;

[]spawn{
    _query = "SELECT ''";    
    _check = [_query, 2] call OES_fnc_asyncCall;    

    o_localTest = _check select 0; 
    o_localChecking = [
        o_localTest isEqualTo "", 
        toArray o_localTest, 
        typeName o_localTest, 
        str o_localTest,
        count o_localTest
    ]; 
    [o_localTest,{    
        o_remoteExecTest1 = _this;    // Will not show if the issue arises
    }] remoteExec ["call",2]; 

    ["",{    
        o_remoteExecTest2 = _this;    // Will always show
    }] remoteExec ["call",2]; 
};

[full sonnet]

I don't know how to reproduce that though

I know of two possible variants of empty strings, and I tested both

[full sonnet]

I would assume that that the issue stems from how empty strings are handled within the dev branch has changed, and that the addon we have based on intercept is having some type of issue with how the data is accessed natively?
It did in previous dev build. But then it was fixed in current one.

[full sonnet]

If there is any way I can potentially scrape memory values/create dumps for this

If you can do it in diag binary (I don't know if Intercept works there)
Run your repro code.
Cheat code CRASH, and click yes to do the fulldump.

Then I can dig into missionNamespace and find it the variable.

Actually also works on normal binary, you can just create the dump via task manager

[mossy cloak]

Anything you want me to change in the code I have to make it easier to find?

[full sonnet]

no

Well

you could put it also in a local variable

and add a while true sleep loop at the end

Then I can just look at the local vars in the spawned script, instead of sifting through mission namespace

[mossy cloak]

Dump created and sent to your DMs with updated code used

[dusk hinge]

TBH I dident have any suggestion, I was moe curious what the plan for it was, but reviewing the wiki updates those are some very nice changes! Love the custom EHs for adding and removing them. I made a EW system I havent pushed out yet, I will defiantly use this for that. I'm manually managing this with the pylonsChanged EH currently.

Not related to pylons but for aircraft still, me and Peral would love the stealth config properties to me changable in the run time. We have quite a lot of stealth aircraft which I made a simple system to enable and disable external pylons but we cant change the stealth characteristics of them. (Im aware this is why the black wasp has two varients)

radarTarget    = 1;
radarTargetSize = 0.35;
visualTarget = 1;
visualTargetSize = 0.7;
irTarget = 1;
irTargetSize = 0.75;

[full sonnet]

Well turns out, there are 3 possible types of strings.

An empty string with a buffer but no content (It used to have content, but was modified in-place and cut down)
An empty string with no buffer (that's what all empty strings are supposed to be)
An empty string with a buffer of size 1, which only contains a null terminator. (The game itself never creates something like this)

[full sonnet]

Oh right, yes someone asked to increase the target sizes based on pylons. I forgot about that.

Mh implementing that is quite annoying and hurts performance (probably not enough to matter, but enough for me to not like it)
Does it really make that much of a difference?

Is it enough to change it by adding a pylon.
Something like "Attaching this pylon increases the size by N"
Or do you need it fully changeable by just script. Just script is probably actually even easier to do.. but still annoying and the performance hit thing

[viscid onyx]

Scripting control over target size would be really helpful for anything that wants to do electronic warfare/countermeasures/advanced radar stuff

Top-of-the-head examples: an ECM pylon that can decrease your radar target size; a scripted system that increases IR target size when firing a laser; vehicles with mode switches between exposed/stealthy configurations

*That last one could even be applied to existing stealth aircraft, by increasing radar size when the internal bays are open

[slender axle]

Yeah this would be awesome. Target size, and maybe even something on the emitter side (radar range sky/ground)

[floral arrow]

Hey guys. RC is resetting my profile. I know it might not happen to all players, but it's still a problem. Happened on DEV as well, but the problem disappeared, but now it came back on RC. Any idea how is this happening and what steps can I do to avoid this? Basically every time I run the game, settings are reset. Ticket: http://feedback.arma3.com/view.php?id=27514 (dev branch, but the problem is the same on RC)

And sometimes now I cannot get trough the "Best place is to start at Bootcamp", if I click no, it changes the Intel part of the main menu, if I click yes it's either crashes or opens up the bootcamp mission selection with an error message.

[formal dome]

Can anyone tell me what the Steam app_update number is for the RC branch. Thx

[tulip bane]

it's an interesting thing that few people seem to report every now and then

btw is here anyone else that has this problem? https://forums.bistudio.com/topic/187971-game-updater-doesnt-load-rc-or-legacy-branch/ I can't download RC (and legacy) branch with Game Updater. Dev and main branches are downloaded without issues.

[formal dome]

I tried app_update 107410 -beta rc validate but I'm still getting 1.54

[formal dome]

Tried: app_update 107410 -beta rc -betapassword Arma3Update157RC validate
and app_update 107410 -beta development -betapassword Arma3Update157RC validate
and app_update 107410 -beta rc validate
none seem to get 1.57.134516

[tulip bane]

isn't it 156 and not 157?

[formal dome]

does it have a password even?

[tulip bane]

yes

[formal dome]

just trying app_update 107410 -beta development validate

hmm

will try

[tulip bane]

app_update 107410 -beta rc -betapassword Arma3Update156RC should be right is I've understood

[formal dome]

rgr trying that

[formal dome]

no did not work. I got 1.56.0.134536 I should have got 1.57.0.134.516

app_update 107410 -beta rc -betapassword Arma3Update157RC does not work either

Got it

app_update 107410 -beta development validate

is correct

[dusk hinge]

How are you implementing it? (Im curious :D)

TBH idk if its something that would make a huge difference but for stealth aircraft with "beast" modes, like the F35 it would be nice. I work with peral on his aircraft and we have two F35s I've setup to have a button setup to enable or disable exterior pylons so for my purpose I would only need a scripting command to change the values.

I think it is going to make negligible difference per weapon on a pylon, but have or not having multiple external pylons would make some difference.

How I assumed it would work is instead of a value being pulled from config those values are stored in variables and can be modified. The only performance hit I could see is when a vehicles sensors checks "can I see that", it now has to pull a variable from each object rather than is config value which is faster. Ofcourse I have no knowledge on the backend or engine workings of arma because its not accessible so im curious how you are thinking to go about this 🙂

Like NikkoJT said too with scripting commands for all of these it will lead to some fun extra things like increasing the IRSize when you fire your afterburner 😄

[full sonnet]

The problem is that these values are currently stored on the entity type. Which is shared between all vehicles of the same type.
So changing it per-vehicle, doesn't work. Thats the annoying thing.

Need to move it over to per entity, but then we have the problem that it uses memory on every entity. Even the ones that are never targets.
Can move it to only transport vehicles, but then when receiving the value we need to check if its a transport vehicle or not, and that costs performance.

[dusk hinge]

RIght, so the per entity is literally EVERY entity. The check for if its a transport vehicle I assume is the same as a isKindOf check? And those loops that need to check for the transport type will only be running on vehicles that have sensor components?

I think the performance drop would be negligible if those are both correct

[tulip bane]

so you wanted dev branch and not rc?

that explains it then 😛

-beta rc is for the release candidate

[past moon]

107410 -beta rc -betapassword Arma3Update156RC

[full sonnet]

Every vehicle that has sensor components, will check almost every other vehicle, so it'll add up

[full sonnet]

https://community.bistudio.com/wiki/setTargetSize

lol..

I noticed an error in my new code. Instead of returning visualTargetSize, I was returning visualTarget, the boolean 0/1 value.
Turns out, that wasn't my error, that was a copy paste.

Its been broken since... probably forever visualTargetSize in config had no effect, it would be either 0 or 1..

Welp lol

[full sonnet]

[] spawn {
  systemChat str diag_tickTime;
  waitUntil [{systemChat str ["check", diag_tickTime]; false}, 10, 2];
  systemChat str diag_tickTime;
}

The interval time only starts after the condition was evaluated, so if the condition itself takes more than one frame, it can take longer than interval between checks.. but this is scheduled so everyone expects it to be messy

[dusk hinge]

ayyy thank you! And yeah I get that, after i wrote my last message I though of that how every vehicle with sensors checks all others.

is this not possible to be global effect? I would of though just on an update of the values they could be broadcasted to all clients as it only would broadcast on update

I can just use remoteExec ofc. I guess there is the argument of what if you want certain vehicles to have "better" radar than others and therefore change target sizes locally

[full sonnet]

is this not possible to be global effect
Yes, but I don't feel like implementing netcode

Can do that later, mayybe

[dusk hinge]

Alright thanks a lot for adding this 🙂

[full sonnet]

I... have a problem here.
I actually cannot use game time for waitUntil timeout/interval, nor ui time....

Due to engine architecture, waitUntil code doesn't have accesss to time.
The only solution I see is to use real world time, not game time.

Which.. should anyway be equal to ui time.
It just means you can't rely on it, and if you really need time (which anyway.. its scheduled, so it'll never be exact) you'd have to keep track of time yourself

[halcyon moth]

Makes it pretty useless for me :/

I think every time I use a timeout in a waituntil, it needs to be simulation time, not UI/real time.

[full sonnet]

I can make a copy of waitUntil that lives elsewhere... but I don't know what'll all break then

[halcyon moth]

Bit baffled that the engine doesn't have general access to simulation time. Like how do you handle AIs getting stuck :P

[full sonnet]

AI's have general access to simulation time

Just this code is in a place that doesn't

[halcyon moth]

Oh well. It was a nice-to-have rather than replacing something really obnoxious.

[full sonnet]

I'll try

[halcyon moth]

On that note I should probably write up an _array findLowest { code }

[full sonnet]

This is ugly right?

_par = createHashMap;
_par set ["name", "testfile"];
_par set ["folder", "subfolder"];
_par set ["content", "R3voman"];

do3DENAction ["CreateTextFile", _par];

This might be better, but I still find it ugly.

do3DENAction ["CreateTextFile", createHashMapFromArray [
  ["name", "testfile"],
  ["folder", "subfolder"],
  ["content", "R3voman"]
]];

I just felt like PoC'ing around a bit..

do3DENAction ["CreateTextFile", #{
  "name": "testfile",
  "folder": "subfolder",
  "content": "R3voman"
}];

This would be the first addition to SQF core syntax, in like.. 20 years?
So I'm a bit scared of doing this.. It seems to work..

# is an indicator for the compiler.
{} is just a normal codeblock
The code block gets compiled into a constant.
And because of the # it inserts a new make hashmap instruction (like make array works in the assembly)
The new instruction, creates a hashmap, stores it like switch statement does its condition, and executes the codeblock, just like switch/do.
The : are alt syntax of the colon command for switch/case, that look up the hashmap and insert into it.

Also available in cursed as F.

_f = {
  hint "hello";
  "name": "testfile";
  "folder": "subfolder";
  "content": getPos player;
};

#{
  call _f
}

Currently it runs unscheduled. So this would also be alternative to isNil {codeblock}. You don't need to put anything into the hashmap.
But I think thats a bad idea, it should support scheduled

[carmine juniper]

Huh, this can help quite a bit of mass config generation etc

[pallid cliff]

Interesting

I have never been a super huge fan of hashmaps as arguments/parameters as they are quite wordy. But that new syntax seems nice.

[teal snow]

Should I make this a ticket

[full sonnet]

Yes. But I'd prefer to not have to do it because its annoying 😄

[teal snow]

There are so many useful, functional applications of having this work

[teal snow]

https://feedback.bistudio.com/T197908

[analog fractal]

Can that write description.ext?

[pallid cliff]

Only text files if you are talking about the new 3den action

[harsh orbit]

@ instead of #?

[full sonnet]

ITs a hash map

🤷

[analog fractal]

well, description.ext is a text file

[pallid cliff]

Literally only .txt file.

[bleak canopy]

All the sensor mast vehicles could use this

[analog fractal]

🙁

[slender axle]

You can almost do this with preprocesser macros but native syntax is cool.
Is there any possibility of adding Sets? Can already accomplish similar with HashMap structure backing but would be nice bridge between Arrays and HashMaps.

[full sonnet]

no

[broken musk]

any rough guidance on which Q we're likely to see 2.22?

[pallid cliff]

He said next release is "monthish" away

[deep sleet]

Wow, this looks like a great feature, but for some mod developers like me, it could actually cause serious problems.

To keep development efficient, we rely on tools like sqflint (a VS Code linter/parser for SQF) to validate our SQF code style and syntax.

If there are syntax-level changes, they could break compatibility with tools like sqflint.

For us, that would likely create more trouble than benefits.

[harsh orbit]

from what I see there's no syntax change to the SQF itself. it's just that now we have alternative syntaxes for some commands (# is now unary with code arg, and : takes ANY instead of CODE)

[deep sleet]

Actually, there is one — for example, : is supposed to be valid only inside a case statement, but now it is also treated as valid in general contexts.

[harsh orbit]

I don't remember there was a constraint on using :. so you can't just write cast "X" : {}; (without any switch) ? it should be possible afaik

[deep sleet]

There is, actually — using : outside case leads to parse errors.
I did try to patch sqflint, but I’m still not very familiar with JavaCC and related tooling.

[harsh orbit]

using : outside case leads to parse errors.
well rn it should because again : is a binary command that takes a CASE object on the left so you do need a case. (so syntax is CASE : CODE)
but once this gets released (if at all) I think Dedmen will change it to be ANY : ANY and the linter shouldn't complain anymore.
external linters like SQFLint will have to be updated manually, but internal linters like ADT will adapt automatically because they can read the syntax data directly from the game

[deep sleet]

Based on my limited understanding, this is currently how sqflint handles : .
It doesn’t treat : as a binary operator/command; instead, it validates : specifically while parsing switch-case statements.

Maybe : could be handled as a normal binary operator instead — I’m not completely sure. If that approach is possible, I can try to work on it, but to be honest, it would still be quite difficult.

We would need to modify sqflint itself to support this properly.
So I just wanted to kindly ask whether this change is truly necessary.
Either way, I understand your reasoning and decisions.

[full sonnet]

Afaik HEMTT is the most up-to-date linting tool.
And they would update to support it

[full sonnet]

using : outside case leads to parse errors.
no it does not. Well Arma does not.

[full sonnet]

#{ is actually a new compiler syntax thing.
Its like a new quotation mark. # is not a command there

So I just wanted to kindly ask whether this change is truly necessary.
You don't have to use it, if your tools are not compatible with it.
What you do in your code is up to you. Why should a old tool that is not being updated, prevent everyone else from getting access to a new feature?

If you were to go by that logic, then any new script command should not be done, because sqflint's command list won't know it and won't know what to do with it?

[unreal lava]

honestly this ^

i'd rather have to keep updating my code to accomodate for engine changes/additions than be "alone" in modding a game where the devs abandoned it without support (having done the latter myself, it's not fun or motivating, spending 50+ hours inexperienced digging through assembly to fix/add/change one feature in the engine because the game was too limiting)

i'd much rather have it keep evolving and spend the 1-2 hours needed if i maybe have to change something on my end because we get new features than to be stuck with a game that'll never grow past its current state, im happy we're still getting updates after more than a decade

[deep sleet]

Okey fine I really appreciate all the work you’re doing, and I’m genuinely glad Arma is still evolving after all these years.
I agree with your overall point — continued engine updates are important, even if they sometimes require extra adjustments on our side.
For me this will likely mean some painful tooling changes, but I’ll adapt and update my workflow as needed.
Thanks again for the explanations, and for continuing to push the engine forward.

[vernal remnant]

@full sonnet Hi i was wondering if you know about the Cycle Order bug where when cycle redos the waypoints after a while they end up defaulting to world origin 0, 0, 0

it appears pretty consistent and I've had to stop using it for my waypoints for ground units, I've tested it with just CBA.

[halcyon moth]

What does "after a while" mean exactly?

[vernal remnant]

A few cycle restarts around 4-5

You'll see waypoints at 0, 0, 0

[halcyon moth]

What type of vehicle?

[vernal remnant]

any could be infantry as well

this is with CSAT.

maybe i have bad code but i dont think so all im issuing is a cycle with waypoints so 🤷‍♂️

[halcyon moth]

If you can, create a minimal replication case: A mission as small as possible that can reliably generate the problem without mods. Then post that to the feedback tracker.

[full sonnet]

no

[tall pumice]

i wonder if the visualTargetSize fix will have an actual influence on the gameplay now or if nobody will notice anything

[unreal lava]

Wooo!

awww 🙁

• Removed: lineIntersects / lineIntersectsSurfaces multithreading (unresolved crashes, though the script syntax is still present)

bi giveth, and bi taketh away - oh well

kinda curious as to how deep of a code edge case nightmare that must've been to cause crashes (i've used it quite a lot actually and never encountered a crash)

[full sonnet]

I have also never encountered a crash while I was testing with it.
I put it in TFAR, none of the players in my unit had issues with it.
But after TFAR update got released to public after the beta testing phase, within one day people started reporting mass crashing.

Still no idea what the issue actually is

[unreal lava]

curious... maybe one day it's figured out 🤷

[halcyon moth]

Ah shit, I just used it in our latest release :P

[full sonnet]

It just iterates the loop on one thread.
Probably still faster than looping it in SQF

[halcyon moth]

Yeah but it's multithreaded on stable?

[full sonnet]

on 2.20 yea

[halcyon moth]

Maybe it's a safe case anyway because it's a lot of 10m rays in a short area.

I would guess TFAR is doing very long rays that might force terrain loading/unloading?

[full sonnet]

TFAR used it for object interception, which is 60m tops

[unreal lava]

it's shrimple, dont need to fix the crash issue, just use gpu for raycasting

(probably changes it just enough to maybe avoid whatever was causing the crash, and totally doesn't include an additional age of work or whatever)

[slender axle]

generic pylons
Are vanilla assets going to uptake these? Would be nice if mission scripting allows to add sensor pods or pilot cameras to helos and jets.

[viscid onyx]

The generic pylon magazines don't require any changes on the aircraft itself to work, and through scripting, any pylon magazine can be added to any aircraft that has pylons. So vanilla aircraft, like all other aircraft, will support generic pylons by default. You will still need a mod to actually create new generic pylon classes to use; I don't think there are any plans to add vanilla sensor/camera pod pylons, because it would require asset work which there aren't resources for.

[slender axle]

Ah I see. Still cool.

And there are fuel pods that got quietly added last patch. Fingers crossed for test sensor / pilot camera pods anyway.

[carmine juniper]

I am not 100% sure about this topic, but is still there a chance of tweaking vanilla P3Ds in the future update? Is torturing reyhard still worth it?

I just actually thought of... Ghillie Suit but with actual head instead of goofy balaclava

Or Bohemian heads don't have facial bones, etc...

[full sonnet]

Small bugfixes yes, but new features rather unlikely

[full sonnet]

I implemented it now, but I don't have something to test it with.

In the sensor class next to "animDirection" is "animPosition", works the same way.
Its all just copy paste and making it position instead of direction so it should probably work...

[teal snow]

Il test with the next update, Thank you

[tawny cedar]

I saw a recent screenshot that showed the A149 without the fuel tank/tgp. Was that just texture/animation trickerybof some sort?

[tawny cedar]

Can this accept coordinate offsets? Or memory point only?

[teal snow]

if its the same as "animDirection", it will be mempoint only

[tawny cedar]

Does the Camera/Sensor pylon support adding a magazine? For example, to let a targeting pod add both the sensors, as well as a laser designator? Or would that have to be done using an EH to add/remove a laser designator weapon/magazine?

[tawny cedar]

So some feedback on sensor/camera pod configuration - I've tested using pod models from a few different mods, since there obviously aren't any in Vanilla. Camera pods work great without any memory point defined (model center default), which allows for backwards compatibility with old models; however, they need to be made with no internal textures if you aren't manually defining a memorypoint, or else you will likely get an obstructed view.

For sensor pods, it seems that the default direction is backwards. Fig. 1 shows a combined sensor/camera pod, and even with PilotCameraRotX/Y attempted for sensor direction (provided by CameraComponent according to https://community.bistudio.com/wiki/Arma_3:_Vehicle_Loadouts#CameraComponent), the associated sensor on the sensor display does not move when the camera is rotated, and constantly points backwards. Fig. 2 shows a sensor pod with no camera conmponent. No default direction is defined, and it points backwards as well.

Personally, I think it might be a good idea to take the LITENING on the A-149 and make a version of the model configured as a camera/sensor pod, just so there's a Vanilla template with all the memory points and directions to allow model makers a reference for what is required on custom pod models.

[full sonnet]

It has script eventhandlers.
You can do anything, even add a weapon with magazines like weapon pylons do, and then use user animation sources to animate the pylon.

[full sonnet]

For sensor pods, it seems that the default direction is backwards.
I fixed backwards direction for sensors like 5 times already
The pod's are actually inverted coordinate system, because they are proxies. The pylonInfo script command is also broken in 2.20 and gives backwards direction
I probably missed something there when the memory point isn't found

[unreal lava]

The missile knows where it is because it doesnt know where its facing

[raven mist]

https://www.youtube.com/watch?v=NEyCfBb9peY

[gray quarry]

wat

[tawny cedar]

For anyone working on TGPs, here's a couple little files I wrote to add/remove laser designators to your aircraft when a generic pylon is added/removed. v1.0 is bugged on two-seaters, fix with switch to addWeapon/MagazineTurret.

        class EventHandlers{
            added = "call compileScript ['\yourFilePath\fn_AL_pylonLaserAdded.sqf'];";//params ["_vehicle", "_pylonConfigClass", "_pylonIndex"]
            removed = "call compileScript ['\yourFilePath\fn_AL_pylonLaserRemoved.sqf'];";//params ["_vehicle", "_pylonConfigClass", "_pylonIndex"]
        };

[tawny cedar]

Some more feedback on camera component: from my test, the Camera Feed display compnent appears to use the memorypoint/direction of the vehicle class' pilot camera instead of the camera pod's properties. Curiously, it does correctly match the vision mode.
Fig. 1: The active pilot camera is underneath the aircraft, with the underside and a pylon-mounted weapon visible. Meanwhile, the Driver feed is taken from the vehicle class' default position with the aircraft's nose and forward terrain visible. It also obeys zoom level of pylon TGPs.

[waxen knot]

...did something get broken with the delete keyword?
Vanilla static AA turrets (Defender/Rhea) and their radars (AN/MPQ-53/Cronus) can somehow be flipped up by a human player if they get knocked over even though their configs specifically remove the UserAction for it (PressXToFlipTheThing).

[tawny cedar]

When assigned by default as an attachment entry to a pylon, generic pylons do not appear as selected in the Attributes->Pylons menu. However it is selectable and may be mounted like normal. Yes, I've double-checked that the version I attached is the compatible one.

[full sonnet]

I think its only the radar display that's rendering it wrongly
This is a radar facing forward, the target display works, it detects the target infront. Only the cone is displayed wrong.

But the weird thing is, if I use a targeting pod with a camera and animation set up, the cone is forwards and also animates correctly

Oh..

XD Everything is wrong

[tawny cedar]

I also dont know if its like this for all models or just pylons (im new to object builder), but I that reminded me i had to invert the Z coordinate of my pylon's pilot camera memory point

[full sonnet]

Pylons are proxies, so they are inverted models.. kinda mess

[carmine juniper]

Inverting the invert, what an invention

[full sonnet]

It is supposed to be, that the direction you set up inside the pylon model, should match.

So if it goes out the front of the pylon model, it should also go out the front ingame.
The engine should handle all the inversions.

I just need to rip out this code and redo it.
The method to get the sensor direction, takes a entity transformation as argument (which seems to be nonsense, because it always gets passed the "same" thing, why even have that as argument..)
Some places pass the transformation of the vehicle having the sensor.
Some places pass the transformation of the owner of the sensor, which so far used to also be the vehicle.

But since pylons have sensors, the pylon is the owner, and the pylons position is inverted and offset.
But some places now pass the helicopter instead of the pylon... ahhh.

[full sonnet]

Afaik there shouldn't be any problems with the pilot camera though

[full sonnet]

There is a bug with the pilot camera creation though.
If the pylon is on there at mission start, the camera position is initialized before the pylon's position, and camera ends up inside the vehicle.
And if the model wasn't loaded yet, the init is delayed, but the pylon offset is applied twice so its widely off too.

[full sonnet]

And the inversion of the proxy was not properly applied to the camera pos, that was probably the thing with your inverted Z coordinate.
Instead of moving forwards it moved backwarsds

[full sonnet]

Man the camera view in the HUD is such a ultra mega pain
Its all based on a memory point on the vehicle.
But its not on the vehicle for pylons

[full sonnet]

Pylon camera position memory point cannot be animated.
I could make it so, but I'd rather not add the overhead

Also HUD camera works by being attached to the vehicle's memory point.
The vehicle doesn't have the pylon memory point, and attaching to pylon is not possible.
But I can attach to the vehicle, with a fixed position offset (So again, the mempoint can't animate here either)
Ah the pain of trying to add a feature to this engine 😄

[harsh orbit]

I wasn't following the conversation. what are you adding again?! 😅

[lucid totem]

goats

[unreal lava]

A horrible one is enough

[lucid totem]

FIRE!!

[tawny cedar]

Ill get some more testing in today/this weekend, one of the camera pylons im adding has a horizonal offset and i want 2 see if that will be inverted

[tawny cedar]

I'm also having this bug (I'm not sure if it's related to Devbranch or not) where if an aircraft doesn't have a particular sensor by default, it won't be able to use that sensor even if it's added by a pylon ammo.

For example, the A-10 doesn't have IR detection capabilities without a TGP or AGM-65 mounted on a pylon IRL, so I removed the VIS/IR sensors from my version of the A-164, with the intent that a TGP or Macer/Maverick could be added to grant the functionality. However, when flying around some VIS/IR targets, none of my weapons, whether it's the cannon or AA/AG missiles, can select a target. Only if I add the TGP to a pylon will it let me do so. Sure for the gun, it's fine, but for the missiles with a built-in sensor, it's definently not what I woiuld expect for functionality. Same goes for IR missiles.

[tawny cedar]

Just tested: Yes, the X-offset of my pilot camera memorypoint is also +/- inverted. Only the vertical/Y direction offset is Accept As Is.

[full sonnet]

The fixes for the stuff should land mid next week

[full sonnet]

I don't know how that works, I'd need a repro to test it

[tawny cedar]

I guess the best way to tell without a custom addon or class would be with the vanilla AH-9
>Place AH-9, assorted air/gnd targets in 3DEN; ensure ground targets have engines on via MOVE waypoint
>Set pylons to Falchion-22, Scalpel, or any other sensor-guided weapon
>Play scenario, attempt to lock on to any target with pylon-mounted sensor-guided weapon. You will be unable to unless you very precisely "boresight-lock" by lining up the target in the exact center of your screen and pressing T/R
It appears unrelated to the new Pylon feature though

[viscid onyx]

If the aircraft doesn't have a usable sensor, it uses the weapon's own sensors. If the weapon's sensors have a very narrow lock cone, then that's what you'll get. You should check the sensor config for those ammo types.

[halcyon moth]

There's one CUP Hellfire with no lock params at all, because apparently they didn't test that case :/

[tawny cedar]

The BIM-9X and Macer lock cones are a lot wider than center-screen boresight though. If a target is selected via boresight, it also only has a target lock diamond, and not a target select square - from this I can presume the Square part of the target symbology refers to aircraft sensor lock and the diamond refers to ammo sensor lock.
Fig. 1: Target is demonstrated to be a valid IR target, and was selected via "boresight" (T button while HUD weapon reticle is centered on target). Only has Ammo Lock symbology because aircraft lacks sensor.
Fig. 2: Same target is well within ammo lock cone (Macer is 35° x 35° IIRC), I am spamming the R button; I could do either R or T to the same lack of effect.
Fig. 3: Same editor file, just have a Generic Pylon added with sensors+camera. Now the target appears on my scope, and I can select it from any angle (i.e. not boresight-locked). If the aircraft was appropriately using the ammo's sensor, this case would be possible without the TGP. The only difference I would expect is the the target to only appear while within the ammo's specific sensor bounds, and the ammo-only lock symbology being present, like in Fig. 1.

No DLCs, just the mod I wrote including this custom derivative of the A-143, sensor pod, etc. It does not change vanilla behavior, and the ammo class is the vanilla Macer.

Similar case, A2A. Heat-seeking missiles like the AIM-9 mounted (vanilla Falchion-22 ammo class) can and are intended to lock IRL with no sensor input due to using passive sensors. However, without an IR sensor on the aircraft or a RADAR activated, my heaters are nigh-useless and unable to lock at close range with very little angular deviation from boresight. It's like if it was only functional using the RADAR slave modes found on real aircraft.

[unreal lava]

dev branch got updated

[lucid totem]

base game too
660 MB

[unreal lava]

am i smelling the 2.22?

[viscid onyx]

Are you sure that's not the EF update?

[unreal lava]

oh

[lucid totem]

perhaps

[unreal lava]

yeah that explains it

[full sonnet]

This weeks dev-branch update is being skipped. So no pylon fixes yet 🫂

[full sonnet]

Same editor file, just have a Generic Pylon added with sensors+camera. Now the target appears on my scope, and I can select it from any angle (i.e. not boresight-locked).
That sounds like the pylon sensors are working fine though..
I don't know what I'm supposed to do here.

The missiles own sensor doesn't work as designed? That's unrelated to the new pylon feature, so its low priority

Which config entry on the missile, decides its lock-on cone?

Is it that the weapon is missconfigured (not my thing to fix), or that the engine doesn't properly handle the weapons configuration?

[broken musk]

little Q about the radio channel stuff... does "setPlayerVoNVolume" hold thru death/respawn? Also does Direct Chat have a hardcoded range?

[tawny cedar]

Im thinking its a separate unrelated engine issue, but only got noticed when I started converting some aircraft but to use pylon sensors.

[full sonnet]

The VoN volume is actually stored on the unit. So I expect no.

Direct chat is 80 meters.
But the volume falloff over it will also make you stop hearing sooner than that

[full sonnet]

Then feedback tracker ticket with precise repo.
I need to know the config properties that set the missile's cone, to know what I'm looking for

[broken musk]

Sweet thanks

[full sonnet]

Over the years (Literally going back to 2013) there were a few requests for verifying DLL Extensions before loading them.
Either by checking their Hash or Signature.
We had an internal proposal (Also 2013) to make "bisign"'s for extensions, but that is not backwards compatible.

There were proposals to add new script command/syntax like callExtension, but you provide a hash to it and it verifies it before loading the Extension. But the problem is that if the Extension was already loaded, using the old syntax, then its already too late.

My idea would be a EventHandler, which you can return false in to block a Extension from being loaded.
This also has the problem of an Extension being loaded before the Eventhandler is added.

So on top of the Mission EventHandler added by script, also have a config class with basically config EventHandlers that do the same, which would be present before the first Extension is loaded, and can do the validation.

The EventHandler would get arguments like allExtensionInfo is providing.
Signature (full certificate chain, only windows), file hash, filesize, modfolder its in, file version info (windows).. What else?

Thoughts?

[harsh orbit]

scripted solution is bad

We had an internal proposal (Also 2013) to make "bisign"'s for extensions, but that is not backwards compatible.
why not?

[full sonnet]

Old Extensions don't have bisigns. They will suddenly stop working.
Server owners would have to make their own, but what for existing workshop mods.
Also once an Extension is loaded, its too late. And you can load an Extension in Singleplayer, before joining a server.

[solid island]

It would make all the existing extensions invalid until modders update them I think

[solid island]

I've never written an extension so I don't know how it works, but does loading it inherently allow the extension to run some malicious code? Or does it require a callExtension to be run first? Could it be an option to block that command before the extension is validated somehow?

[full sonnet]

Just loading it is enough for it to do anything

[harsh orbit]

Also once an Extension is loaded, its too late. And you can load an Extension in Singleplayer, before joining a server.
well yeah. this is true of any solution you provide. no?

[full sonnet]

A config class that validates it.
The server has no control over the Extension. But the mod that provides it does, and the mod is bisign checked, and you cannot unload a mod.

[harsh orbit]

if you load a mod that server doesn't use and the mod is not signed you can't join the server at all right? sounds like a trivial question but I've never done it so idk

[full sonnet]

Yes

[ebon quail]

depends on signature validation setting of the server no? but in this case I assume it would be set to verify all

[solid island]

Is it more to protect the players from loading "unverified" dll extensions, or rather make sure that the dll versions match between the server and the client?

[full sonnet]

The main focus is "unverified".
But if the DLL is verified and you can trust that the DLL is what it says it is, then you can also implement the version match check in script.

Currently you can implement version check.
But you trust that what the DLL tells you is its version, is actually it.
In the age of open source extensions, everyone could build their own and replace it with one that just states the same version.

[solid island]

Maybe backwards compatibility is not so bad if you prompt the user to confirm they want to load a mod which has a missing dll signature

[full sonnet]

Prompt doesn't work, we cannot delay the DLL loading if its requested, without breaking backwards compat

[pallid cliff]

So what if the mod with the DLL gets updated. The hash of the dll would change so you need to push an update to check against the new hash?

[full sonnet]

Use signatures. Same private/public key, no need to adjust

[pallid cliff]

rgr

[full sonnet]

Currently allExtensions has sha1 hash, does it actually matter enough that I should also provide sha256?
Because of sha1 being old and "insecure" and stuff

[full sonnet]

(I know exitWith doesn't work properly.. please ignore that fact)

addMissionEventHandler ["ExtensionLoad", {
  if (_this get "name" != "ace_break_linecorrupt") exitWith { true }; // return, Allow loading

  // must have signature, so signatures can't be empty
  if ((_this get "signatures") isNotEqualTo []) exitWith { false; };

  // Find a signature that is..
  _validSignatureFound = (_this get "signatures") findIf {
    _x get "publicKey" == "308201...E4EA81730203010001" // Signed by my key
    &&
    !("expired" in (_x get "status")) // Not expired
    &&
    !("badDigest" in (_x get "status")) // The signature actually matches the binary (it was not edited after signing)
  };

  _validSignatureFound != -1 // If signature found we load, otherwise we block
}];

Argument is a hashmap

{
   "name":"ace_break_line",
   "hash":"676CE410E48A0D3CED71004D4F1FAC11D99D0D55",
   "signatures":[
      {
         "notAfter":"2026-04-12T16:42:56Z",
         "serialNumber":"1022",
         "extensions":{
            "authKeyId":"DB479B6C8A1F05807EDBAFCBC2823DB52D554E9C",
            "comment":"Plugin Certificate",
            "subjectKeyId":"92CF8B549101560DD5D771754E44326EE4056935"
         },
         "issuerName":"DE, Germany, Intercept, CA, Intermediary CA Dedmen",
         "issuer":{...},
         "status":[
            "ok",
            "expired",
            "revocationUnknown",
            "partialChain"
         ],
         "notBefore":"2023-04-13T16:42:56Z",
         "thumbprint":"B6C731E2ABE2113723FAB0EC5342292B576B0E27",
         "publicKey":"3082010A...203010001",
         "subjectName":"DE, Germany, Intercept, CA, Intercept Core Extension",
         "counterSignSig":{...}
      }
   ],
   "filesize":226136,
   "productVersion":"3.8.4.0",
   "hashsha256":"FFB537D2D08B081E9AB387C629F6223D00E6A6149EA91AB45E255D7B6CCDBDA7",
   "fileVersion":"3.8.4.0",
   "mod":"curator"
}

[unreal lava]

had this conversation in another chat (i think it was in scripting) - not really a failsafe per se as much as a warning (against malicious ones at least), but if there was a prompt when starting a game without battleye whilst using battleye-unverified extensions (or any extension in general if you cant check for battleye verification) that essentially tells you "you're using an mod xyz that contains an extension that wasnt verified by battleye, if you have been instructed to do so on purpose, it may contain malicious code/etc..."

(i know it's not exactly on topic but since its an extension conversation i figured i'd drop my two cents here)

[full sonnet]

that wasnt verified by battleye
We don't know that

[unreal lava]

gotcha, figured as much but wanted to be sure - i mean then it's just along the lines of "using mods with extensions without battleye is dangerous"

can have a "dont tell me again" cause im sure many people do this, but at least it's a warning that still happened

[verbal venture]

Would be funny if BE was also just checking the hash of the DLL

[full sonnet]

For their whitelisting, they basically are yes

Find a hash collision in whatever hash they use and have fun

[full sonnet]

Where to put the config eventhandler to verify the extension.

Its basically a MissionEventhandler but in config

class CfgExtensions {
  class MyModName {
    extensions[] = {"extension1", "extension2"}; // Whitelist, which extension names to run this check on (so its not ran on unrelated extensions we don't care about)
    onCheckLoad = "call MyMod_fnc_checkExtensionLoad";
    publicKey = "308201...E4EA81730203010001"; // The handler has access to the config class, so we can store some properties here)
  }
}

Everything i come up with seems ugly in some way.

MyMod_fnc_checkExtensionLoad = {
  // must have signature, so signatures can't be empty ("signatures" entry is not present on linux")
  if ("signatures" in _this && ((_this get "signatures") isNotEqualTo [])) exitWith { false; };

  // Find a signature that is..
  _validSignatureFound = (_this getOrDefault ["signatures", []]) findIf {
    (_x get "publicKey") == (_this get "config" >> "publicKey") // Signed by my key (Easy access to config that added the handler, so can store the key there)
    &&
    !("expired" in (_x get "status")) // Not expired
    &&
    !("badDigest" in (_x get "status")) // The signature actually matches the binary (it was not edited after signing)
  };

  _validSignatureFound != -1 // If signature found we load, otherwise we block
}

[raven mist]

To be fair though that provides a lot of new info and looks like it would do the job :)
-# Maybe @mossy cloak might have some thoughts on #dev_rc_branch message though?

[full sonnet]

Well I have that in for next profiling branch update.
The config one only, not the missionEH one.

Might even keep it config only, not decided yet

[full sonnet]

We actually have a ticket about that from 2013. Back then it would've been viable to do that without breaking backwards compat. But 13 years later.. not so much.

We could make a "Hey this mod contains extensions in its folder, we don't know if it will try to load them, or if they are even able to load, but beware"
But... the Arma Launcher already has that.

There was a proposal of moving extensions out of the game process, into a separate process.
Overhead for communications over there, but security benefit. And we can just kill/restart the process when we join a server. So no extensions can load before, and we could actually apply bikey/bisign verification there.

But 13 years since that proposal..
And it would've also blocked some recent features, like interacting with graphics.

[gray quarry]

why it wasn't done back then?

[lucid totem]

apparently, some other minor developments happened in between really

[gray quarry]

hm

[nova wigeon]

maybe stupid question(maybe not) what happens when i use callExtension in that onCheckLoad callback 🤔 does that crash the game? does it just not load the dll and throw an error?

[teal snow]

did this end up in the perf version or the next dev version?

[full sonnet]

Ah right I forgot that.
It'll be ignored if that is also to an extension that's not yet loaded.
It will fail/skip the load.

Also if a script blocks an extension, it will stay blocked and not retry. Which I also forgot.

[full sonnet]

That should be in Prof.

[teal snow]

Iv just tested Prof, this does not appear to work, I have the animPosition = "sensorPos"; set inside my ActiveRadarSensorComponent components, I do not have an "animDirection" defined and it still appears to be originating from model centre.

[full sonnet]

I'd need something to test it myself

[teal snow]

Iv got something but it aint simple, I don't know of a way to do this simply, I cant config patch a mempoint into a vanilla vehicle to test

Might have an easier way

[teal snow]

Sent you a PM to test it

[compact salmon]

Has anybody actually gotten the "required mods" thing working in 1.56RC server browser?

http://imgur.com/G5WlR5u

http://imgur.com/HOEkA2B

For me it just doesn't seem to able to connect "@CBA_A3" running on the server with the "@CBA_A3" I've got from steamworkshop

[compact salmon]

@past moon Any idea what I'm doing wrong?

[unreal cypress]

Do you have CBA downloaded from steam as well, or have you downloaded it from somewhere else and added it manually?

[broken musk]

any chance an old mission maker vet could get a personal favour added to the game?

request: description.ext option to auto-join players into a randomly selected lobby slot (instead of the current first-available selection)

it would address this issue:

right now if there’s a bugged slot (appears joinable while it’s actually occupied), all joining players get auto-joined to it and get “stuck on receiving data”

if slot selection was randomized, the likelihood of encountering the bugged slot would be way lower

joinRandomLobbySlot = 1

please ❤️

then i can finally use the “skipLobby” option and bypass the lobby UI altogether. at the moment it’s essential to have so players can select a non-bugged slot

[raven mist]

To bump on this we actually have had this issue forever too. I've never brought it up because I've never been able to reproduce or get a dump to provide useful enough info.

We recently had a case where 8 or so slots went at once and people couldn't join them. Upon debugging I found the names associated with the units were either null or nil:

[swat_0, swat_1, swat_2, swat_3, swat_4, swat_5, swat_6, swat_7] apply {getPosATL _x}
// [[0,0,0],[0,0,0],array,array,[0,0,0],array,array,array]

Typically it'll only happen with a few slots at a time so we are more convinced the one I was debugging above is mission related. Those specific slots were also only implemented within the year to support adding a faction to a gamemode; and what I did to make it work wss quite hacky. It does make me wonder though if in instances of this happening where we've always thought it an Arma bug that somehow the unit is being deleted.

[full sonnet]

Easy. Feedback tracker ticket please

Does it need some extra config through? Like random joining across all factions?
Only across blufor, until its full and then bump into redfor until its full and so on?
For you to think about and plan out

[full sonnet]

nil. getPos doesn't return nil on null arg

[tepid dock]

Added... Latin?
Is this a "because we can" case, here or was there a special reason behind this one?

I'm asking because it's been a long time since I last met someone speaking it

[lucid totem]

a special request by @pontifex in DM

[winter crescent]

https://www.bohemia.net/blog/arma-2-the-name-tale

Arma is originally a Latin word
(No special reason other than it was a high-level request from inside BI)

[pallid cliff]

*These languages were experimentally translated using an AI model. The quality of these translations is not guaranteed, and they may contain mistakes. Separate issues are unlikely to be fixed, but full regenerations using improved models may occur. All other existing languages were translated by a team of human translators.

May I ask what AI model was used?

[ebon quail]

GPT-5.4
But subject to change for whatever reason

[pallid cliff]

I wonder how that compares to DeepL

[tall pumice]

oh no, what's vojtěch thinking about this

[pallid cliff]

Pulling his hair out

[halcyon moth]

On the nil/null, IIRC player vars return nil until someone has actually spawned in a slot. If they joined and then left then it's objNull. Bit of a pain.

[gray quarry]

was latin ai translated because there's nobody at bi's reach who knows latin translation?

[teal snow]

I recently generate string tables for my mod using ai translations, the feedback iv had so far was okay with minor contextual errors in the translation

[raven mist]

Same- I figure an AI translation is better than no translation at all. I can't control that I don't speak French, German, Spanish but I can try to do what I can to support it.

[teal snow]

exactly, and if anyone does speak the language and wants to correct it, Il welcome that

its immedietly significanty easier for people to enter using mods if there is somewhat of a translation instead of nothing

[tepid dock]

Depends. A botched translation may be more annoying, because it could break the immersion every now and then. It would constantly remind me "oh, right, I'm just playing a game, because those lines don't make sense in this context so it's not a real person who said that", so to speak

[raven mist]

I guess I could see that in the case of mission making with AI who speak dialog. I feel like for other stuff like UI I'd still rather use a botched translation over a language I don't understand though.

[broken musk]

https://feedback.bistudio.com/T198797 For my uses I only have blufor configured. I think @trim crane had the same bugged lobby slot issue but he was using the lobby to select factions so players were manually selecting their slot

    0,          // EAST, random role-joining disabled
    1,          // WEST, random role joining enabled
    1,          // RESISTANCE, random role joining enabled
    0           // CIVILIAN, random role joining disabled
};```

obviously excluded module/logic/zeus slots so players cant shuffle themselves into special roles

[trim crane]

Maybe the array should be side index instead?

[raven mist]

A little off-topic from the current request but pretty sure when they're locked on receiving data they also can't escape to chose a new slot. It'd be cool if maybe there was some sort of timeout or something before allowing user input to escape but that's probably way easier asked than to do.

Right now only real option is to ALT+F4 and not sure if we'd be able to use the randomized slot selection since our blurfor slots have units that require different whitelists. So maybe even just some way to chose which slots are viable too if it's not too large of an ask 👉👈

[fervent remnant]

Create zeus slots and control the world

[carmine juniper]

Well machine translation was ever introduced to the game all the way back then, like 8 years ago. I'm not against of these new strings but it's not something entirely new. Our people suffered by playing Bootcamp already, before we fix it

[unreal arrow]

That order in Quicksilver's example is how the sides enum ordered

[trim crane]

assignRandomRoleForSides[] = {1,2};
```I think this looks neater

[broken musk]

are players sorted evenly by faction when they first join the server @trim crane ? i’ve only ever worked with 1 lobby faction

my intention is to use the skipLobby option. its unusable if there’s a bugged slot that every client is getting assigned to, so random assignment will sidestep the bug

[misty aspen]

Imo having numbers in a true/false config looks funky but I get the idea

What could be neat is replacing the idea with the faction name

So

assignRandomRoleForSides[] = {east,west};

[tall pumice]

cant you do that with #define

[misty aspen]

Why make complicated when can be simple

[pallid cliff]

Gotta be a string then.

Just stick to numbers and use defines as Lexx said.

[misty aspen]

Fair

[frozen pumice]

Is there any chance to make https://feedback.bistudio.com/T63330 ? Request from 2013

[unreal arrow]

Doesnt it convert to string anyway?

[unreal arrow]

What happens if you use the same folder?

[frozen pumice]

All servers have same list of missions

[unreal arrow]

Right so what is the problem?

[full sonnet]

I have it on my list to change it. To set the var to objNull by default, instead of nil.
But, it could break scripts if people do isNil check to check if slots are occupied or not

[ebon quail]

So you can define the list of mission in their config with missionWhitelist and/or missions class?
And the config is already different per server due to name etc, so you already manage multiple server.cfg's

[frozen pumice]

Technically I can, but it's too complicated for my case. The first server (stable missions) must have one list of missions that must be changed every few days, and another server (for testing (pre-play and validate)) with another list of missions that must be changed too. So, I have two folders with different missions for that

[unreal arrow]

I'll have a look

[broken musk]

along those lines, I've always wanted to be able to use just one PBO for all terrains... the only thing holding us back is the mission.sqm file

imagine if the system would look for fileExists (format ["sqm\%1\mission.sqm",worldname])

or search for fileExists (format ['%1.sqm',worldname])

and if found, run that instead of mission.sqm

this would mean a unified PBO could handle all terrains

[broken musk]

on current dev branch, getting a client crash when I open the "Direct Connect" tab in the server browser

[ebon quail]

already listed on known issues :D
is fixed already, just not in that build

[fervent remnant]

mission.sqm file

#include "defines.sqf"

#ifdef __USE_LINGOR 
#include "sqm\lingor.sqm"
#endif

#ifdef __USE_SAHRANI 
#include "sqm\sahrani.sqm"
#endif

#ifdef __USE_CHERNO 
#include "sqm\cherno.sqm"
#endif

#ifdef __USE_CHERNO_W 
#include "sqm\cherno_w.sqm"
#endif

#ifdef __USE_Caribou 
#include "sqm\Caribou.sqm"
#endif
...

[broken musk]

hmm, where is __USE_terrain defined?

and does this actually work?

[fervent remnant]

#include "defines.sqf"

[broken musk]

can i see the defines.sqf?

just trying to set up a PBO which works on all vanilla terrains

[fervent remnant]

I made this decision back on arma 2 and it worked fine.

[fervent remnant]

defines.sqf

//#define __USE_TAKI
//#define __USE_SAHRANI   // привязка к местности на карте
#define __USE_LINGOR   // привязка к местности на карте
//#define  __USE_Panthera
//#define __USE_Taki
//#define __USE_Fallujah
//#define __USE_CHERNO_W
//#define __USE_Esbekistan
//#define __USE_CHERNO   // привязка к местности на карте
//#define __USE_Caribou
//#define __USE_isladuala
//#define __USE_Celle2
//#define __USE_Podagorsk
//#define __USE_Emita
//#define __USE_Utes
//#define __USE_Napf
//#define __USE_Franken
//#define __USE_Pantera
//#define __USE_Kulima
//#define __USE_Everon
//#define __USE_Malden

[broken musk]

thanks mate

[misty aspen]

hm thats interesting, the file still needs the .mapname though or? so you cant just do a wildcard pbo that shows up for every map

[fervent remnant]

and then I remembered what the problem was... Yes, every mission had to have a map name in the pbo file.

[misty aspen]

Ah yeah sad times

[gray quarry]

where can the known issues list be found at?

[gray quarry]

ah so it's put in the changelog on a per patch basis as it gets found? thought there might be a dedicated place for that
thanks

[ebon quail]

there is one generic one (its probably out of date)
https://dev.arma3.com/known-issues

but for dev we just list the dev branch specific ones that are currently a problem in the changelog for that specific dev branch release

its probably out of date
like this one which doesn't exist anymore and just redirects to some ford raptor wallpapers site

[gray quarry]

heh

who manages that page?

[ebon quail]

[pallid cliff]

Obviously no one. Time to purge it and link to the changelog category on the wiki.

[gray quarry]

very funny

[pallid cliff]

Was dead serious though.

[ebon quail]

changelog ones are for dev branch known issues, outside of these.
like "dedicated servers crash on startup" and stuff like that

[pallid cliff]

We don't have a biki page for stable branch issues?

[gray quarry]

hm fair enough, sorry for misassuming

[gray quarry]

no, that page fullfills that role for general known issues with the released/stable version

[halcyon moth]

Doesn't even have an AI category :P

[unreal arrow]

So it will be -mpmissions= param and it will only work on dedicated server

[frozen pumice]

That's perfect

[fervent remnant]

Yes, I would really like to be able to enter the missionname empty world (missionname.empty) into the pbo file and then upload the desired world inside the file.

[misty aspen]

Yee

[unreal arrow]

So format is -mpmissions=MPMissionsAlt or -mpmissions=MPMissions\Experimental but no absolute paths like in -profiles=

[gray quarry]

relative to game directory only? no dot wildcards such as ..?

[unreal arrow]

nope, mpmissions is designed to be relative to arma3

[tall pumice]

who broke the tigris hud :> i deployed 2 smoke grenades, but the monitor hud still shows 2 green squares

[winter crescent]

I make a very shallow pass as part of every main branch release, but those are quite infrequent these days. Could probably do a deeper clean-up next time.
(Unlinked Raptr, thanks for the spot.)

[ebon quail]

can probably remove it entirely, seeing how it stopped existing in 2017

[winter crescent]

But could someone still have it somehow locally installed? 🤷‍♂️ Will mark to remove next pass.

[waxen knot]

Uhh...were the Repair Depot props from the Tanks DLC always like this?
DestructBuilding seems to be glitching out for whatever reason.

[ebon quail]

all of you gotta stop giving me heart attacks with these random issues after we repacked the entire game data(currently on dev) 😂
seems to be a code issue, not data luckily

[gray quarry]

sure but then people wouldn't know about that issue if they forgot about having that still installed

[gray quarry]

imo keep it for historic value since it's technically not fixed by the discontinuation of the app

[ebon quail]

There is no relevance to keep a mention of an online/social app thats been defunct/down since 2017. thats a skill issue on having it installed at this point 😂

thats checks calendar
almost 10 years ago

[gray quarry]

8 years ago*

i still have a defunct gaming hub app installed that im lazy to remove

[ebon quail]

Its also not just it being installed its having the client open/running

[gray quarry]

ah fair enough

[gray quarry]

ah thanks for your service /gen

[pallid cliff]

Why are you still featuring Arma 3 Units in the launcher then?😬

[ebon quail]

don't worry, 0 issues today 😌

[pallid cliff]

Still more reliable than GitHub ATM.

[gray quarry]

why does it have so many issues?

[gray quarry]

nobody wants to deal with the launcher code spaghetti
iirc dedmen is in charge of launcher code?

[raven mist]

I had an effing heart attack yesterday. The night before I merged a bunch of shit into our main branch when I was super tried. Imagine my surprise when I saw no pull requests in our repo the next morning. I thought I somehow screwed up default branch

[halcyon moth]

Oh hey, does that on perf too :P

I remember those are weird because the live one shows up in vehicles but the dead one doesn't appear in allDead.

[broken musk]

this looks like the repair depot "healthy model" is clashing with the code which tries to keep supply/ammo containers from sinking below terrain

healthy model set underground, wreck spawned, code to protect supply containers from sinking keeps pushing it back above ground, clashing with the ruin model

solution probably for the code which protects from sinking to check for a ruin model

[nova wigeon]

*no issues until now(i would not ask for it that way tbh 🤣)

[misty aspen]

and discord seemingly

[full sonnet]

Doable. Make a FT ticket please
How to find what missions work on the terrain though. missionName.*.pbo ?
But what if a SQM is then missing.. But I think likely doesn't matter anyway, we already open the pbo and read mission.sqm to get the mission name out of it. We can read pbo and see what sqm's are inside.

[full sonnet]

Solvable too. Preprocessor could have predefined macro with world name. But, I prefer multiple SQM filenames like Quiksilver proposed.

[full sonnet]

Trying to find green squares in the image, but failing to

[tall pumice]

those should turn red when firing smoke launchers. it works in other vehicles

[full sonnet]

bla bla #dev_rc_branch message

[full sonnet]

Already kinda broken in 2.20 stable. You can see it wobble about. Just not as badly, it still stays mostly inside the wreck.
In prof v50 its already wobbly
v40 is wobbly
v31 is wobbly
v10 is not
v18 is not
v24 is wobbly
v22 is not
v23 is not.

#arma3_branch_changelog message

Fixed: Some ThingX entities would get suspended too early when they are un-alived (Shooting target plates not falling down)
ugh..

[tall pumice]

add more weight to the object?

that said, does it really need to be thingx? it's not like you can flip it back into a box and transport it somewhere else

oof, it already has 500000 mass 😄 maybe it's too much? my mortar container scenery has 7000 mass and uses a similar config

looks like it gets destroyed similar to buildings. maybe just swap it to house simulation directly?

[broken musk]

maybe it just searches for worldname.sqm ? otherwise falls back to mission.sqm behaviou

[full sonnet]

It uses DestructBuilding. And we already have a bypass for DestructBuilding if the destruction phase is 0.
But this thing is sinking into ground, so destroy phase starts at 3 and goes up. Its not zero.

[full sonnet]

Doable. But I don't want to search all PBO's.
So the pbo itself needs to be marked to support multiple terrains

• is not allowed in filename.

The problem is yes one pbo can support multiple terrains. But then you need 50 copies of the same pbo, because the game only loads pbo's with terrainname.pbo

[broken musk]

yep ill leave those ideas up to you. the idea is just to have a universal mission file which works on multiple terrains.... ill make the FT now

[full sonnet]

Yeah maybe I just make it use mission.sqm as fallback.
And then you might load into a unknown terrain you never heard of.. and I'll let you deal with that problem..
Or we support mission.sqm fallback not existing, and make the mission just not show up in the list then

[broken musk]

https://feedback.bistudio.com/T198959

how does the engine decide which terrain to load? is it from pbo filename or mission.sqm

[full sonnet]

pbo filename. The mission.sqm has no info about what terrain it is

[broken musk]

this is just what i had in mind

[full sonnet]

mission.stratis.sqm is what I have in mind

[broken musk]

nice

while you are here, any tips on how to run a clone of bis_fnc_arsenal ? i want to use the display/HUD and adjust some things under the hood, but running the display is causing trouble

[full sonnet]

No

Trouble is there to be had, not to be avoided

[broken musk]

hahaha understood

[full sonnet]

If you have an idea to configure which slots (list of indices could work), then put it on the ticket.
I'm pretty sure the top slot is just 0, and then so on. Per side.

[full sonnet]

Using enums isn't really anything new.

If configuration becomes more complicated, with indices per side, and maybe priorities which side to fill first or whatever, it might also make sense to just have class Blufor opfor, ...

[full sonnet]

But if you have separate server configs, why command line parameter. Wouldn't it make sense to have the mission folder in config.
So you only need to swap config, and not config AND mpmissions parameter?

[full sonnet]

Yeah.. just do that in all the 54 places where we try to check for mission.sqm ez.

[frozen pumice]

Hm... idk. In my case there is no difference for me: command line or config. Command line argument is more intuitive, maybe

[full sonnet]

Command line has additional complications. Not for us but for users.
Like anyone using a server hoster, nitrado and such. Do they generally offer a "extra command line parameters" or do they only allow you to set what they offer? They likely would take a long time to support new parameters.
While server.cfg is generally freely editable and you can directly use the new features.

I think in most cases, like yours (and like Crumble brought up), you'd anyway also want to change the server name to identify which mission set is active. If you use a mission rotation, you have to change it too.
So I think in most cases, you would end up always having to change two parameters instead of just one.
Mission folder to me seems like server configuration, and server configuration seems to make sense to be in the server configuration file 😄

Also the fact that we now have a parameter, that only works if you're a server. (And afaik that's the only one that does that. Well -config parameter is also server only)
And server config already only works if you're a server.

But now is kinda too late, as the parameter is already implemented now

[pallid cliff]

Could use a Param file, that way hoster doesn't need to support new parameter.

[ebon quail]

they would still have to support the param file arg, which they probably dont

[gray quarry]

it's not too late to undo the implementation but whatever

[misty aspen]

In my experience there's usually a field for custom args but not sure if every hoster has it

[halcyon moth]

I don't remember any hoster having a field for custom args. It would be pretty dangerous for them. We found an exploit to add arbitrary args on Armahosts and they did not appreciate it :P

The Linux ones could maybe allow it because they're probably using Docker.

[misty aspen]

ope fair

[full sonnet]

configfile >> "CfgVehicles" >> "O_APC_Tracked_02_AA_F" >> "MFD" >> "MFD_Commander_Smoke_Indicator"
I can see the text to the left of it.
But can't find anything that controls the thingies being green. On my screen it just looks white too 😄

[tall pumice]

it's model.cfg issue

well, at least i am suspecting it is. actually, i'll check

[full sonnet]

"cmAmmo" would be the source for it

I can see that source being used in airplanes and helicopters. But not in any tank MFD

If it is model.cfg, then its data.
And in 2.20 data I can't see it working either

[tall pumice]

ok, first things first, it works fine in the kamysh, which inherits from exactly the same stuff. except that there is z-fighting. alpha sorting would need to be refreshed, or draw index needs to be increased to "high" in model file

i remember now that i fixed this when doing the vehicle variant for the ws 1.1 update

what it should look like:

[full sonnet]

Tigris also doesn't work on 2.20 stable. So not a new issue

[tall pumice]

config side looks correct to me. it inherits everything from the base class as it should. animationsource is there, it's also the same smoke launcher weapon

[full sonnet]

model.cfg has two "mfd smoke" selections. That "hide" based on "Smoke_source" source
Tigris has "Smoke_source" as "revolving" on weapon "SmokeLauncher"

[tall pumice]

yea but that should be ok

[full sonnet]

Kamysh just inherits the same. And uses the same base model.cfg for that

[tall pumice]

the only real difference i see is that mfd_com_smk_x selections in tigris model is also in OctoVez selection. and it is also tied to the same bone "mfd_com_smk_1", "otocvez", "mfd_com_smk_2", "otocvez",

this could cause the rotation to fuck up

what i would try is to remove the OctoVez selection of those two mesh in the p3d file and try again

[full sonnet]

I don't know how that works.
If its a model.cfg issue, its 6 years old and I don't care right now.

[tall pumice]

it could be broken since tanks dlc added it, because at around the same time, the kamysh damage model was also broken :p (and fixed later)

two issues in model.cfg: "drivewheel", "", bone is duplicated, and in line 486 a ; is missing after "speed"

hah, i was right. the additional OctoVez selection is causing it

[full sonnet]

I'll make a ticket for reyhard 🤷

[tall pumice]

it can also be fixed with just removing the octovez link from the tigris skeleton: #dev_rc_branch message
but i would suggest to fix it in model and setting z bias to "high" to also fix the flickering issue. z bias can also be set to high on the kamysh model, to fix it there as well. ez, takes 20 seconds including opening the model and saving the changes : >

[broken musk]

just revisiting an old request @full sonnet , is an AI/group "AttackTargetChanged" event handler still not feasible?

[full sonnet]

If it wasn't feasible back then it still isn't.
If it just was that I didn't have time to look into it back then, then it might be different now

[full sonnet]

Is that offer still up? Cuz REST API is planned for after the next update 😄

[torpid sand]

lol. I no longer dev for the server I was with that could’ve used it. But I’m sure they’ll be happy to have it. @mossy cloak

[lucid totem]

an http one, not an https :p

[full sonnet]

[full sonnet]

There is still nothing on the FT ticket.
And the ticket also seems unclear.

So if east is disabled, but west is enabled.
If west is full, you jsut get the first east slot, no randomization?
Random side selection is also not a thing?

I also considered for other cases anyway, to make a server side script eventhandler, that lets you specify in which slot a joining player should go by default.
Can also be used to assign players based on their playerID, to specific slots.
You can implement randomziation however you want in script, more configurable than I could ever make it with a config file.

if I make it scripted, then all the config stuff is obsolete.

[fervent remnant]

Why not simply:
Create a virtual spectator slot
Set skipLobby = 1 in description.ext
Then handle it however you like...

[full sonnet]

Because thats nonsense

[misty aspen]

arma 3/BI branded sponsorship fursuit

[waxen pumice]

Where can I find more information on this feature?

Tweaked: Third-party on-screen action available in Spectrum Device view is now executable

I am not aware of it.

[tall pumice]

navid lod5 is missing camo selections 🥲

[grim lion]

Http REST API? Please tell me more I'm listening... Lol

[unreal arrow]

equip spectrum device, come to a vehicle, get in action appears, press space, it works now

[full sonnet]

So you can now use normal actions while holding a spectrum device?
That changelog entry is a bit unclear then 😄

[waxen pumice]

@tame wraith FYI
#dev_rc_branch message
This might effect CrowsEW because the default keybind is also used for recording bearing lines on the map.

[unreal arrow]

If you are near vehicle and getin shows then it will trigger the get in, so either get in and do it while inside or step away from vehicle

[viscid onyx]

You can also open the action menu (e.g. scroll) and then close it again (e.g. rightclick) to temporarily dismiss the screen-centre action. I think that will also stop the "currently selected action" button from activating it.

[unreal arrow]

No the scroll is and was always disabled when you use spectrum device

I just realised we have ability to detect the scrolldown menu if it is visible but no idea if an action is shown on screen

[tame wraith]

Thanks! Might just need to change the default key. The key is already configurable for players iirc.

[frozen pumice]

smth strange, when I'm using -mpmissions with relative path, it logs an error and falling back to default mission

arma3server_x64.exe -mpmissions=Servers\VTG\mpmissions ...

19:55:31 Can't find mission template: '\mpmissions\VTG_New_Paralel_147_v3.vt7' in 'Servers\VTG\mpmissions\', skipping to next mission.

looks like it passing relative path instead of just filename while looking in -mpmissions dir

[unreal arrow]

what directory your normal MPMissions is?

[frozen pumice]

just root

like default one

[unreal arrow]

and servers?

[frozen pumice]

"Servers" is just folder that contains profiles in root directory

[unreal arrow]

Arma3
.. MPMissions
.. Servers
.. .. VTG
.. .. .. mpmissions
this should work

[frozen pumice]

yes I have exactly same

but it doesnt

[unreal arrow]

It probably made the directory for you already, it creates it if doesnt exist check in Arma3 dir there should be Servers

What is your server config

template=

[frozen pumice]

just one mission, to make it default instead of lobby

[unreal arrow]

That doesnt have VTG_New_Paralel_147_v3.vt7

[frozen pumice]

yes, it doesn't. Should it be here?
without -mpmissions all mission listed in <arma>/mpmissions work

[unreal arrow]

Where is it now what config?

why is server looking for VTG_New_Paralel_147_v3.vt7 if it is not in config?

[frozen pumice]

Was selected via #missions

[unreal arrow]

do you start it with -mission param or something?

[frozen pumice]

Nope

#missions shows all missions from -mpmissions directory, but they cannot be selected

[unreal arrow]

I havent checked if #missions is supported for the new folder option

so if you type #missions what is showing

[frozen pumice]

all missions from MPMissions + from directory specified in -mpmissions

[unreal arrow]

Aere you in the lobby when executing #missions?

lets make it simple, could you write down in the ticket detailed repro so I can check properly

[frozen pumice]

sure

thank you

[frozen pumice]

https://feedback.bistudio.com/T199168

[unreal arrow]

Thanks I will have a look

[unreal arrow]

Could you do me a favour and tell me if #missions works with say -mpmissions=mycustommissions

[frozen pumice]

just 1-level folder? yes, 5 mins

[unreal arrow]

yes

[frozen pumice]

nope, still error
Can't find mission template: 'sions\beach.Malden' in 'mycustommissions', skipping to next mission.

#mission <missionName> works btw

[unreal arrow]

right right, and if you just rename it to custom?
-mpmissions=custom

[frozen pumice]

looks like I said before, lobby selector tries to set mission by full path (truncated)

[frozen pumice]

5m

[unreal arrow]

there is some parsing bug that works with short name but fails on longer names

[frozen pumice]

works with custom

[unreal arrow]

yeah I will look tomo too late now, it is folder name for some reason cannot be too long, the mission name can

0:56:09 Mission T197336_blahblahblahblahblahblahblah.VR read from bank.

Yep, -mpmissions=cus\tom also works

there is bug in parsing pretty sure I know why it does this

[frozen pumice]

if found another issue related to this, but not sure if it fixable on your side.
all missions added by this option are not listed in BE Rcon

rcon shows missions from MPMissions only

but #mission <name> works via RCon as well

even if mission not listed in RCon's missions

should I create another task on bugtracker?

[unreal arrow]

Yes, I will discuss this with @full sonnet afterwards

[frozen pumice]

https://feedback.bistudio.com/T199178

[unreal arrow]

there was hardcoded length of the mpmissions folder

[gray quarry]

oop

[waxen pumice]

I just tested the new behaviour on dev-branch.
It is not just spacebar that is affected. It is also the middle mouse click that now often triggers the vanilla action menu and might close the Spectrum Analyzer when one wants to zoom in.

@unreal arrow also some more feedback in
https://feedback.bistudio.com/T198138#2946414
and the post right after it.

[unreal arrow]

You zoom in with RMB no?

[tame wraith]

I believe he means our mods custom function that "zooms in" on the highlighted spectrum. (Essential changing the em_min and em_max to a smaller range). So we might need to rethink using that key when in the spectrum device display 😅

[unreal arrow]

gotcha, just to clarify this only happens when you have another action displayed right?

[tame wraith]

@waxen pumice ^
(I have not tested it myself yet)

[waxen pumice]

No, RMB is open/close Spectrum Analyzer

• Zoom is Middle-Mouse-Button.
• Shift+Middle-Mouse-Button is unzoom to factory default.

[waxen pumice]

Correct

Not sure if this changed behaviour would also affect the Contact DLC gameplay.

[unreal arrow]

when spectrum is opened now it zooms in if you hold RMB

it wont affect Contact because Contact uses Contact's assets

[waxen pumice]

How does it open and close the Spectrum Analyzer then? 🤔

[unreal arrow]

you press RMB it shows big display, you press and hold it it zooms in just like scope in weapon

[waxen pumice]

Oh wait. I guess we are talking about 2x different things.
What I mean with "zoom in" is what @tame wraith said here #dev_rc_branch message

Zooming in in the spectrum range.

[unreal arrow]

yeah I got that

[waxen pumice]

@tame wraith am I right that your input scheme is directly taken from the Contact DLC single player behaviour (I guess for ease of familiarization)?

[unreal arrow]

\a3\UI_F_Contact\UI\RscInGameUI these are Contact's scripts

[waxen pumice]

Yeah. I forgot that it has a separate (albeit very similar) script.

[unreal arrow]

vanilla uses a3\UI_F_Enoch\UI\RscInGameUI

the scripts were just copied not sure if anything was altered in the process

[waxen pumice]

I compared them (on 2.20). They only show minor differences.

[unreal arrow]

I will have a look if I can hide those action icons instead of enabling them

I also added shownAction script command that tells you if action icon is shown on screen

[waxen pumice]

Oh that would be awesome, because they were always confusing, even if they were never triggered. 🙂 👍

[tame wraith]

I actually don't recall. I think I chose middle-mouse because nothing else was using it when zoomed in and it seemed familiar for that type of "selection" tasks.

[waxen pumice]

That doesn't make a difference. Because it is not just the default action (space bar) that is now triggering. It is also the vanilla menu entry that is highlighted (when middle-mouse is triggered).

[tame wraith]

does the vanilla action-menu show up when you scroll the middle-mouse now? Or does the spectrum display take control of the mouse scrolling as usual?

[unreal arrow]

no scroll is still disabled

middle mouse button is also Action just as spacebar

[waxen pumice]

No, the vanilla menu shows when pressing middle-mouse, not when scrolling. Scrolling still works.

[tame wraith]

yeah okay. So you can get that menu to show, but I assume when its shown you still can't navigate it with scroll like usual until you close the spectrum device?
So that can cause a bit of confusion, even ignoring our mod.

but yeah, we can find another key to use instead if needed

[waxen pumice]

@tame wraith in 2.20 (current stable) the vanilla menu also shows when you press middle-mouse-button (which is confusing enough).
But the action is never triggered when pressing middle-mouse again. It keeps zooming in into the spectrum with every middle-mouse button click.

In 2.21 some middle-mouse presses will zoom in into the spectrum and some will cause a vanilla menu action to trigger (if it was shown).
Workaround is to always cancel the vanilla menu with RMB every time middle-mouse spectum zoom was used.

Maybe CrowsEW using middle-mouse just needs changing to Ctrl+Middle-Mouse to sort things out (would somewhat match Shift+Middle-Mouse for the spectrum unzoom).

[tame wraith]

ah gotcha. Yeah that is not ideal.

Yeah, changing the default key is probably easiest.

Assuming the vanilla menu is not triggered when using combination of keys even if it includes middle-mouse.

[waxen pumice]

Only problem is that even Ctrl+Middle-Mouse let's the vanilla actions take notice of the middle-mouse press.

Ctrl+Numpad+ and Ctrl+Numpad- seem to still be available for something like spectral zoom in and out.

Because I totally see that KKs change to make the actions work, while the Analyzer is open, makes perfect sense.

[tame wraith]

yeah okay. Hmm, makes it more difficult. I kinda wanted to be able to do it without removing your hand from the mouse. As it would often be something you would do on the fly while aiming around scanning/tracking etc.

But I guess changing the default to something non-conflicting, and then allowing players to rebind it as a standard CBA keybind might be the way to go. Then each can figure out what works for them.

[waxen pumice]

Ctrl+Mouse-Up and Ctrl+Mouse-Down would also be free as it seems.
That would even be in line how a lot of GUIs handle zoom in and out.

[tame wraith]

That is a good shout

[waxen pumice]

I'll make a test implementation to see how it feels.

[tame wraith]

But might want to redo the "zoom" to be smaller increments centered around the selection, instead of being the entire selection. To be more of a scroll action feel of "zooming"

[waxen pumice]

@unreal arrow responded to your questions https://feedback.bistudio.com/T198138#2946421

[unreal arrow]

So the BE RCon uses own DLL to look into MPMissions directory, so it has no idea if game uses custom directory. The clue is

command starts without # it has BE internal handling
command starts with # it is passed over to Arma 3 and handled there

I could make Arma 3 print out those missions, but, the mecjaninsm that allows Arma 3 to print arbitrary message to RCon is currently broken in the BE DLL. If it gets fixed I will implement something like #mpmissions

[frozen pumice]

that's sad. Ok, so we are waiting for fix from BE side?

[gray quarry]

sounds like a hard ask

[unreal arrow]

I found terrible hack how to do it and I like it

[misty aspen]

based

[unreal arrow]

#mpmissions
Player #0 SERVER (192.168.0.30:2302) connected
(Command) SERVER: mpmissions\
(Command) SERVER: T197336_blahblahblahblahblahblahblah.VR
(Command) SERVER: old.T182166vehicle.Stratis
(Command) SERVER: oldT197336_Dedicated_Server_language_Startup_Parameter_Not_Working.VR
(Command) SERVER: old_restartNoRE_repro.VR
(Command) SERVER: restartNoRE_repro.VR
(Command) SERVER: Silent_Bullet_Test.Altis
(Command) SERVER: T182166vehicle.Stratis
(Command) SERVER: T197336_Dedicated_Server_language_Startup_Parameter_Not_Working.VR
(Command) SERVER: temp_restartNoRE_repro.VR
(Command) SERVER: test_Silent_Bullet_Test.Altis
Player #0 SERVER disconnected```

[gray quarry]

hm

bandaging features at its finest :v

[unreal arrow]

That could change if BE dll gets updated

[gray quarry]

even with the industry credentials, are be people hard to get in touch with?

[waxen pumice]

@tame wraith Ctrl+mouse-wheel works quite nice.
I hope the video captures it.
https://youtu.be/GbWCN6LHTls

Changing CrowsEW to this would avoid the double meaning of the middle-mouse button.

[tame wraith]

That looks great and more intuitive! Definetly an improvement.

[unreal arrow]

I am looking into default actions not showing rather than enabling so bear with me

[full sonnet]

@unreal arrow can we not add a variable and let mods disable this using the action menu thing?
Then crows ew (and maybe other mods?) just needs a one liner fix instead of rethinking their whole hotkey scheme

[unreal arrow]

i have it 95% finished everyone is going to love it

[unreal arrow]

hideActions, hiddenActions... and before you ask, no it is only possible for player

[unreal lava]

Been wanting a way to hide/remove actions for a while

[unreal arrow]

Whoa, it actually works pretty good. hideActions [91] hides all user added actions leaves only engine actions, hideActions [-91] does the opposite

Best feature? hideActions [0] hides ALL actions but it is still possible to execute things like reload R or open inventory I

[raven mist]

I'm a little out of the loop on the action changes- am I understanding what I am reading right that this would allow us to hide default actions?

[unreal arrow]

yeah basically but only for player, you cannot target specific vehicle, because even vehicle actions are collected and added to payer when player looks at a vehicle

[raven mist]

That's huge

We've had dozens of requests over the years to have a setting that would remove the inventory action

Would finally be able to do that

[full sonnet]

So for Crows EW.
If you have a spectrum device in hand, they can just disable all actions, and the "Get in" vehicle one won't show up?

[unreal arrow]

No, I will hide all actions when device is up, anyone can then ovewrite it if they need action enabled. Before action would show but none was executable

[ocean sail]

Can we expect a button in the game settings to hide actions from the actionmenu that have a button on the keyboard?

[unreal arrow]

no you can do it yourself wth the new command

[ocean sail]

in all scenarios? even for example king of the hill

[unreal arrow]

You could request that @trim crane adds it as an option

[trim crane]

#arma3_feedback_tracker message

How I imagined this implemented, through an event handler

Alternatively it could be attached to either target or caller (and called when something wants to show an action on a target or a caller tries to show any action)

player addEventHandler ["HandleEngineActionShow", {
    params ["_action", "_unit", "_target"];
    _action in ["UAVTerminalReleaseConnection", "SwitchToUAVDriver", "SwitchToUAVGunner"]; // true for handled
}];

[unreal arrow]

The actions are collected from all over the place to be added to player, those that are about to get added have another filter where you can hide actions you supply (see shownAction enum), hide all actions but the one you supply, hide all or hide none

no event handlers

[trim crane]

Yeah those events would be called each frame for tons of actions

Performance

Limiting actions just for player might not be ideal either, you often can have an AI do an action for you instead, I assume this implementation will still allow for that

So say you did hideActions [-91] to block everything engine, but then you can ask AI to say open a door that you can't open yourself

Or are config actions also considered user actions? 🤔

[unreal arrow]

So, currently this is pretty safe implementation giving you ability to decide what to show to player. I am not going to go beyond this because 2.22 is too close

[trim crane]

If it is, bad example then. Say you want to block "Start timer" on a satchel charge like that, you still could start it through AI actions, right?

[lucid totem]

if AI is local (and you can give them orders) there is a chance that the local machine disables all actions everywhere, but I'll let KK answer that one 😬

[unreal arrow]

you could execute any action with action command they are unaffected

[unreal arrow]

called it hideActions on purpose because it essentially hides unwanted actions, doesnt disable doesnt interfere with other functionality == safe

[trim crane]

I assume you provide an array of actions you want to disable, right?

[unreal arrow]

yep it is either [1,2,3..] or [-1,-2,-3...] or [0] mix and match results in script error. or []

[trim crane]

Got it, you just showed a single number in your examples so I wasn't sure

I want to disable "UAVTerminalReleaseConnection", "SwitchToUAVDriver", "SwitchToUAVGunner" actions from mouse wheel menu in my mission to unclutter it (you can do all these in UAV terminal dialog)

Being able to supply strings instead of action IDs could be helpful too

Do we even have that action ID list anywhere exposed to us?

[unreal arrow]

hideActions [163, 165, 166 ]

[trim crane]

I still wonder if its possible to implement it using a target/caller argument 🤔

[unreal arrow]

Being able to supply strings instead of action IDs could be helpful too
Dunno how I feel abou tit

The filtering happens in the end, to add what you ask it has to be done in multiple places so no

[trim crane]

player hideActions ["StartTimer"]; // Can't start timer on any satchels
{_x hideActions ["StartTimer"];} foreach units group player; // You can't select the action through AI control menu
_satchel hideActions ["StartTimer"]; // Can't start timer on that particular satchel

[unreal arrow]

gimmie the repro I'll have a look

Just to check, I suspect the units actions collected and added to player and filtering happens at the last moment so they wont show, but better be sure

Do we even have that action ID list anywhere exposed to us?
https://community.bistudio.com/wiki/shownAction

[trim crane]

(Nevermind)

[unreal arrow]

?

[trim crane]

Forgot you have to add the satchel with addOwnedMine to bot for him to have the actions

[ocean sail]

When I ordered subordinate to deactivate mine then SPD VERT %.0f str_ui_speed_vertical_freefall text shown under his icon

[unreal arrow]

still I am curious if it will be axed

[trim crane]

Making a repro script, almost done

[trim crane]

bot = group player createUnit ["B_Soldier_universal_F", player modelToWorld [0,2,0], [], 0, "CAN_COLLIDE"];
[bot] joinSilent group player;
doStop bot;

satchel = createMine ["SatchelCharge_F", player modelToWorld [0,2,0], [], 0];
satchel2 = createMine ["SatchelCharge_F", player modelToWorld [0,10,0], [], 0];

bot addOwnedMine satchel;
player addOwnedMine satchel;

bot addOwnedMine satchel2;
player addOwnedMine satchel2;

player addAction ["hideActions [61];", {
    call compile 'hideActions [61];'; // Can't start timer on any satchels
}, 0, -100, false, true, "", "_this == player"];

player addAction ["[NEW] player hideActions [61]; Forbid start timer for player", {
    call compile 'player hideActions [61];'; // Can't start timer on any satchels
}, 0, -100, false, true, "", "_this == player"];

player addAction ["[NEW] bot hideActions [61]; Forbid start timer for the bot", {
    call compile 'bot hideActions [61];'; // You can't select the action through AI control menu
}, 0, -100, false, true, "", "_this == player"];

player addAction ["[NEW] satchel2 hideActions [61]; Forbid start timer on second satchel specifically", {
    call compile 'satchel2 hideActions [61];'; // Can't start the timer on second satchel
}, 0, -100, false, true, "", "_this == player"];

Actions for current implementation + my suggested one which takes either target or caller

F2 -> 6 to check the timer action

[unreal arrow]

Player yes, bot no because that is different menu, commanding menu. Also https://community.bistudio.com/wiki/hiddenActions
https://community.bistudio.com/wiki/hideActions

[leaden heron]

thats super nice, it would be also really helpful to be able to stop people from certain hotkey actions, right now if people set an hotkey onto a mouse button there is no way to stop it

[unreal arrow]

It hides unwanted actions, doesnt disable them. For example inventory bound to I is an action called Gear. when you are next to vehicle it appears and you can open inventory with either middle mouse button or spacebar. If you hide it, neither middle mouse button no spacebar would open it, but it will still be available via I

If action is bound to a key chances are it will be executable by that key

[analog fractal]

oh neat we can finally hide the artillery computer properly in ACE

[unreal arrow]

the functionality is added to player, if you respawn you will need to readd it

[unreal arrow]

@lucid totem could you please stick WIP on both commands, what I added to wiki might need to be tweaked

[lucid totem]

on hideActions/hiddenActions, sure thing

done

[lament crow]

This has been requested repeatedly over time and now the devs have delivered. Thank you! 🙏

[tame wraith]

That is pretty neat. Could potentially also allow to add "jamming" for base-game explosives "touch-off" action. e.g. hide the action on players entering a jamming radius of effect.
Although might also be smarter ways to deal with that, was just random #showerthought

[unreal arrow]

It is WIP

[unreal arrow]

127.0.0.1
Enter the game server's port:
2304
Enter the RCon password:
123

Logged in successfully

RCon admin #0 (127.0.0.1:63646) logged in
missions
Missions on server:
old.T182166vehicle.Stratis.pbo
oldT197336_Dedicated_Server_language_Startup_Parameter_Not_Worki
temp_restartNoRE_repro.VR.pbo
test_Silent_Bullet_Test.Altis.pbo

#mpmissions
Missions on server (mpmissions\):
MORE\T197336_blahblahblahblahblahblahblah.VR
old.T182166vehicle.Stratis
oldT197336_Dedicated_Server_language_Startup_Parameter_Not_Working.VR
old_restartNoRE_repro.VR
restartNoRE_repro.VR
Silent_Bullet_Test.Altis
T182166vehicle.Stratis
T197336\T197336_Dedicated_Server_language_Startup_Parameter_Not_Working.VR
temp_restartNoRE_repro.VR
test_Silent_Bullet_Test.Altis

BattlEye came through

[unreal arrow]

check out MORE\T197336_blahblahblahblahblahblahblah.VR
this pbo is inside MPMissions\MORE
to launch it
#mission MORE\T197336_blahblahblahblahblahblahblah.VR

[unreal arrow]

Which is best?

hideActions ["HideSelected", 1, 2, 3]
hideActions ["UnhideSelected", 1, 2, 3]
hideActions ["HideAllButSelected"]
hideActions ["UnhideAllButSelected"]

hideActions ["HideSelected", [1, 2, 3]]
hideActions ["UnhideSelected", [1, 2, 3]]
hideActions ["HideAllButSelected", []]
hideActions ["UnhideAllButSelected", []]

"HideSelected" hideActions [1, 2, 3]
"UnhideSelected" hideActions [1, 2, 3]
"HideAllButSelected" hideActions []
"UnhideAllButSelected" hideActions []

[tall pumice]

2nd? i would prefer array style boxed numbers

[unreal lava]

+1 on leftarg/rightarg, seems ideal for two arg commands

[unreal arrow]

@trim crane ?

@lucid totem ?

@harsh orbit ?

[harsh orbit]

2nd I guess?

[raven mist]

Ditto on 2nd

3rd is cool though too

[trim crane]

Other SQF commands often take upper case strings, could fit here too

"CAN_COLLIDE" and such

[lucid totem]

I vote ["string", [values]]
I would use a different name on third otherwise

[trim crane]

Also I think the command should accept readable action names alongside number IDs

[lucid totem]

that's because enum
but yes case-insensitive

[unreal arrow]

It will be case insensitive internally

2nd it is

[trim crane]

@unreal arrow Any chance for contextual hiding by providing target/caller as argument?

[unreal arrow]

Ok, the action has target so theoretically you can provide target and only filter actions added from target, but

if you want to track every target you have to store hideActions for every target, and that is not happening

it is stored in player brain as the actions are displayed for player

[trim crane]

Hmm

Can't you store hidden actions list in the entity itself?

_entity hideActions [61];
```Player checks if the action ID 61 has to be shown, if it has target, check if target has 61 in its hidden actions array

player hideActions [61];
```Player checks if the action ID 61 has to be shown, check if the player itself has 61 in its hidden actions array

[unreal arrow]

no it will add too much complexity, it is alrready becoming quite complex with ability to set reset some actions while leaving other actions for other mods to modify

[gray quarry]

sounds ambitious

[unreal arrow]

This is why I like 2nd syntax, we can expand arguments if needed

[trim crane]

hideActions createHashMapFromArray [
     ["HIDE", [1,2,3]]
    ,["UNHIDE", [4,5,6]]
    ,["HIDE_ALL_EXCEPT", [7]]
    ,["UNHIDE_ALL_EXCEPT", [8]]
];

But yeah I'm OK with 2nd syntax

Still think the actions should be addressed by their name so the code is more readable

[unreal arrow]

if this is possible to implement at a later date we can add it hideActions ["mode", [61], _entity];

[trim crane]

Personally I'd just go the entity oriented approach

This makes the most sense, how often do you need to disable ALL actions? Not sure. How often do you want to disable specific action on that specific entity? Pretty sure that's what people are gonna use the most

[unreal arrow]

I'm not happy with string for mode, it could be #define HIDE_SELECTED 1;
hideActions [HIDE_SELECTED, [1,2,3]]

[trim crane]

Sure you can do defines for presudo enums but its still ugly, the actions have names, it makes sense to let scripters use the names

Not a big deal for me personally but I just think this is right

[trim crane]

We probably could poll people who have a need for this feature and ask for their use cases

[unreal arrow]

I started this command BECAUSE I need to disable all actions

initially it was going to do just this disableAllActions

but then this would be a waste of command with great potential

shownAction shows enum you can always find out what number you need without going to wiki